Project Title : BANK CUSTOMER CHURN ANALYSIS

MY PROFILE

from PIL import Image

img=Image.open('Customer-Churn-1.png')

img

INTRODUCTION TO LOGISTIC REGRESSION

Logistic regresion makes it possible to predict a variable that is dichotomous ,meaning a prediction that requires only two variable.eg a yes or no answer , true or false , big or small, man or woman etc.

The varable to be resolve is called , predicted , target or criterion ,while the varables you need to predict the ###results are,called the independent variables, or the predictors.

Examples of different machine learning logistic regression algorithms are ,

Naive Bayes

Logistic Regression

K-Nearest Neighbours

Support Vector Machine

Decision Tree

Bagging Decision Tree (Ensemble Learning I)

Boosted Decision Tree (Ensemble Learning II)

Random Forest (Ensemble Learning III)

Voting Classification (Ensemble Learning IV)

Neural Network (Deep Learning) etc and we will be using some of them in this project

AIMS AND OBJECTIVES OF CARRYING OUT CUSTOMER CHURN ANALYSIS:

Customer churn is the rate at which customers unsbscribe or leave a company .It can be for a virety of reason like ,

unpleasant customer service , increase in cost beared by customers , un updated technologies and many other reasons.

Customers are inevitably the most important part of any organisation, the lost of a customer simply means reduction in

revenue , which can lead to slow business growth in a company or business and in most cases , constant increase in customer

churn, may lead to outright closure of the company or business.

Hence customer churn analysis and building the right model for a business to avert customer churn ,is a critical and very

important aspect of every company ,knowing the right reason a company customer churn and very early can help the company ,

adress the issue effectively and stop any future customers churn timely.

customer churn also help companies to be able to segment customers in groups reduce marketing expences , by understanding how to target the right potential customers through targeted marketing etc , it also gives companies ,opportunity to understand their customers needs better.

PROBLEM STATEMENT:

An old bank have discovered that , every month, the number of V.I.P , cutomers unsubscribing and leaving their bank monthly , despite the bank reducing intrest rate to encourage customers and also introducing a lot of new automatic bank transaction packages at no extra cost ,to encourage customers to stay with the bank ,have no effect , as more customers are leaving the bank, than the number of new customers subscribing to the bank .

Further preliminary analsis shows the following

1, The cost of getting new customers to subscribe to the bank is 25% more expensive than retaining old customers,

2, Most new customers that subscribe, replacing the V.I.P. customers are average customers with low income ,hence low savings balance in their account. which is generally reducing the total income generated by the bank monthly.

Through another analysis the bank also determined that ,if this customer churn trend continue at this rate monthly , the bank may become bankrupt in 24 months peeriod and needed a fast and permanent solution to these higlighted problems.

The bank have contacted me to determine the following and advice them accordignly:

a, Why are customers churning on the increase monthly ,especially the V.I.P. customers

b,Which insights can be derived from my analysis to highlight the main problems

c, Advice the bank managers or management on how to solve this problems , based on insights derived from the customer transaction data provided to me , by the bank.

d, The bank also want me to build a solution model ,that can be used by the bank to prevent this issue in the future.

BENEFICARIES OF THIS ANALYSIS

1, SALES DEPARTMENT

2, CUSTOMER SERVICE

3,MARKETING DEPARTMENT

4,PRODUCTION : GOODS/SERVICES

5, LOGISTICS

6, COMPANY MANAGEMENT

APPROACH :

1, UNDERSTANDING THE PROBLEM STATEMENT

2, GATHERING OF NECESSARY DATA FOR ANALYSIS AND MODEL DEPLOYMENT

3, STUDYING AND UNDERSTANDING THE DATA

4, CLEANING / WRANGLING THE DATA

5, QUERRYING AND PERFORMING ANALYSIS ON THE DATA WITH VISUALIZATIONS

6, FUTURE ENGINEERING / DATA PREPROCESSING

7, SPLITING THE DATA

8,APPLYING RELEVANT LOGISTIC REGRESSION MACHINE LEARNING ALGORITHMS

9, DETERMINING THE RESULTS BEST ACCURACY

10 BUILDING AND DEPLOYING THE MODEL FOR FUTURE PRODUCTION

1, IMPORTING THE PYTHON LYBRARIES AND DATASET :

import pandas as pd
import pandas as pd
import seaborn as sns
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
import warnings
warnings.filterwarnings("ignore")
import itertools
import missingno as msno
import os
for dirname, _, filenames in os.walk('/kaggle/input'):
    for filename in filenames:
        print(os.path.join(dirname, filename))
pd.options.display.float_format = '{:.2f}'.format
import warnings
warnings.filterwarnings('ignore')        

churn1=pd.read_csv('C:/Users/HP/Downloads/churn_modelling1.csv')
                  

churn1.head()

	RowNumber	CustomerId	Surname	CreditScore	Geography	Gender	Age	Tenure	Balance	NumOfProducts
0	1.00	15634602.00	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00
2	3.00	15619304.00	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

churn2=pd.read_csv('C:/Users/HP/Downloads/churn_modelling2.csv') 

churn2.head()

	Row_Number	Has_CrCard	ISACTIVEMEMBER	Estimated Salary	Exited
0	1.00	1.00	1.00	101348.88	1.00
1	2.00	0.00	1.00	112542.58	0.00
2	3.00	1.00	0.00	113931.57	1.00
3	4.00	0.00	0.00	93826.63	0.00
4	NaN	NaN	NaN	NaN	NaN

COLUMN NAME CONSISTENT

# USEING RENAME FUNCTION TO MAKE ALL COLUMNS NAME CONSISTENT
churn2.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

churn2.rename(columns={'Row_Number':'RowNumber','Has_CrCard':'HasCrCard'},inplace = True)

churn2.rename(columns={'    Estimated  Salary':'  estimated salary'},inplace=True)

churn2.columns

Index(['RowNumber', 'HasCrCard', 'ISACTIVEMEMBER', '  estimated salary',
       'Exited'],
      dtype='object')

churn2.head()

	RowNumber	HasCrCard	ISACTIVEMEMBER	estimated salary	Exited
0	1.00	1.00	1.00	101348.88	1.00
1	2.00	0.00	1.00	112542.58	0.00
2	3.00	1.00	0.00	113931.57	1.00
3	4.00	0.00	0.00	93826.63	0.00
4	NaN	NaN	NaN	NaN	NaN

USE STR REPLACE TO REPLACE SPACES BEFORE OR AFTER THE COLUMN NAME

churn2.columns=churn2.columns.str.replace(' ','')

WE CAN REPLACE COLUMN HEAD WITH THIS METHOD

churn2.columns

Index(['RowNumber', 'HasCrCard', 'ISACTIVEMEMBER', 'estimatedsalary',
       'Exited'],
      dtype='object')

churn2_col=['rownumber', 'hascrcard', 'isactivemember', 'estimatedsalary','exited']

churn2.columns=churn2_col

churn2.head(2)

	rownumber	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	1.00	1.00	101348.88	1.00
1	2.00	0.00	1.00	112542.58	0.00

churn1_cols=['rownumber','customerid','surname','creditscore','geography','gender','age','tenure','balance','numofproducts']

churn1.columns=churn1_cols

churn1.columns

Index(['rownumber', 'customerid', 'surname', 'creditscore', 'geography',
       'gender', 'age', 'tenure', 'balance', 'numofproducts'],
      dtype='object')

churn1.head()

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts
0	1.00	15634602.00	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00
2	3.00	15619304.00	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

churn1=pd.read_csv('C:/Users/HP/Downloads/churn_modelling1.csv',names=churn1_cols,header=0)

churn1.head()

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts
0	1.00	15634602.00	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00
2	3.00	15619304.00	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

ADD THE TWO DATA SET INTO A NEW VARIABLE NAME CHURN

MERGE FUNCTION

churn=pd.merge(churn1,churn2,on='rownumber',how='left')

START TO STUDY THE DATA FOR BETTER UNDERSTANDING

#TO SHOW THE 5 FIRST ROWS OF THE DATASET
churn.head()

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00
2	3.00	15619304.00	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	1.00	0.00	113931.57	1.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	0.00	93826.63	0.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

churn

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00
2	3.00	15619304.00	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	1.00	0.00	113931.57	1.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	0.00	93826.63	0.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
10027	NaN	NaN	Johnstone	516.00	France	Male	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10028	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10029	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10030	9999.00	NaN	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	1.00	0.00	92888.52	1.00
10031	10000.00	15628319.00	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	1.00	0.00	38190.78	0.00

10032 rows × 14 columns

# TO SHOW THE LAST 5 ROWS OF THE DATASET
churn.tail()

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
10027	NaN	NaN	Johnstone	516.00	France	Male	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10028	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10029	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10030	9999.00	NaN	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	1.00	0.00	92888.52	1.00
10031	10000.00	15628319.00	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	1.00	0.00	38190.78	0.00

churn.index

Int64Index([    0,     1,     2,     3,     4,     5,     6,     7,     8,
                9,
            ...
            10022, 10023, 10024, 10025, 10026, 10027, 10028, 10029, 10030,
            10031],
           dtype='int64', length=10032)

churn.sample

<bound method NDFrame.sample of        rownumber  customerid    surname  creditscore geography  gender   age  \
0           1.00 15634602.00   Hargrave       619.00    France  Female 42.00   
1           2.00         NaN       Hill       608.00     Spain  Female 41.00   
2           3.00 15619304.00       Onio       502.00    France  Female 42.00   
3           4.00         NaN       Boni       699.00    France  Female 39.00   
4           5.00         NaN        NaN          NaN       NaN     NaN   NaN   
...          ...         ...        ...          ...       ...     ...   ...   
10027        NaN         NaN  Johnstone       516.00    France    Male 35.00   
10028        NaN         NaN        Liu       709.00    France  Female 36.00   
10029        NaN         NaN        Liu       709.00    France  Female 36.00   
10030    9999.00         NaN  Sabbatini       772.00   Germany    Male 42.00   
10031   10000.00 15628319.00     Walker       792.00    France  Female 28.00   

       tenure   balance  numofproducts  hascrcard  isactivemember  \
0        2.00      0.00           1.00       1.00            1.00   
1        1.00  83807.86           1.00       0.00            1.00   
2        8.00 159660.80           3.00       1.00            0.00   
3        1.00      0.00           2.00       0.00            0.00   
4         NaN       NaN            NaN        NaN             NaN   
...       ...       ...            ...        ...             ...   
10027   10.00  57369.61           1.00        NaN             NaN   
10028    7.00      0.00           1.00        NaN             NaN   
10029    7.00      0.00           1.00        NaN             NaN   
10030    3.00  75075.31           2.00       1.00            0.00   
10031    4.00 130142.79           1.00       1.00            0.00   

       estimatedsalary  exited  
0            101348.88    1.00  
1            112542.58    0.00  
2            113931.57    1.00  
3             93826.63    0.00  
4                  NaN     NaN  
...                ...     ...  
10027              NaN     NaN  
10028              NaN     NaN  
10029              NaN     NaN  
10030         92888.52    1.00  
10031         38190.78    0.00  

[10032 rows x 14 columns]>

churn.sample(frac=0.5)

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
2793	2787.00	15585100.00	Rioux	511.00	Germany	Female	40.00	9.00	124401.60	1.00	1.00	0.00	198814.24	1.00
752	746.00	15640059.00	Smith	606.00	France	Male	40.00	5.00	0.00	2.00	1.00	1.00	70899.27	0.00
3262	3256.00	15671387.00	Fetherstonhaugh	507.00	France	Female	29.00	4.00	89349.47	2.00	0.00	0.00	180626.68	0.00
6666	6650.00	15635277.00	Coates	605.00	Spain	Male	47.00	7.00	142643.54	1.00	1.00	0.00	189310.27	0.00
4668	4662.00	15670416.00	Ferri	780.00	France	Female	43.00	0.00	0.00	1.00	0.00	1.00	15705.27	0.00
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
7524	7508.00	15642001.00	Lorenzen	576.00	Germany	Male	44.00	9.00	119530.52	1.00	1.00	0.00	119056.68	1.00
8007	7991.00	15643635.00	Robertson	664.00	Spain	Male	32.00	5.00	133705.74	1.00	0.00	0.00	134455.84	0.00
9916	9900.00	15811594.00	Gordon	660.00	Spain	Female	28.00	3.00	128929.88	1.00	1.00	1.00	198069.71	0.00
3519	3513.00	NaN	Boylan	806.00	Spain	Male	18.00	3.00	0.00	2.00	1.00	1.00	86994.54	0.00
7472	7456.00	15748499.00	Johnson	550.00	Germany	Male	33.00	4.00	118400.91	1.00	0.00	1.00	13999.64	1.00

5016 rows × 14 columns

churn.sample(7)

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
1970	1964.00	15771139.00	Douglas	578.00	Germany	Male	34.00	8.00	147487.23	2.00	1.00	0.00	66680.77	0.00
3998	3992.00	15601659.00	Fiorentino	496.00	Germany	Female	59.00	7.00	91680.10	2.00	1.00	0.00	163141.18	1.00
4513	4507.00	15635177.00	Williamson	597.00	Spain	Female	66.00	3.00	0.00	1.00	1.00	1.00	70532.53	0.00
8960	8944.00	15727350.00	Pai	516.00	France	Female	37.00	8.00	113143.12	1.00	0.00	0.00	3363.36	0.00
9790	9774.00	15698462.00	Chiu	532.00	France	Male	36.00	4.00	0.00	2.00	1.00	1.00	132798.78	0.00
3666	3660.00	15664668.00	Zarate	534.00	France	Female	42.00	9.00	144801.97	1.00	0.00	1.00	12483.39	1.00
1860	1854.00	15633574.00	Montes	730.00	France	Female	41.00	4.00	167545.32	1.00	1.00	0.00	128246.81	0.00

churn.sample(7,random_state=0)

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
7109	7093.00	15723884.00	Nekrasova	758.00	Spain	Male	40.00	3.00	0.00	2.00	0.00	0.00	96097.65	0.00
3847	3841.00	15662884.00	Naylor	739.00	Germany	Male	58.00	1.00	110597.76	1.00	0.00	1.00	160122.66	1.00
5957	5941.00	15791958.00	Mazzi	849.00	France	Female	41.00	6.00	0.00	2.00	1.00	1.00	169203.51	1.00
6787	6771.00	15815295.00	John	662.00	France	Female	38.00	2.00	96479.81	1.00	1.00	0.00	120259.41	0.00
2849	2843.00	15785782.00	Ugonna	513.00	Spain	Male	48.00	2.00	0.00	1.00	1.00	1.00	114709.13	1.00
3523	3517.00	15787151.00	Liao	638.00	France	Female	34.00	7.00	0.00	2.00	1.00	1.00	198969.78	0.00
6287	6271.00	15702806.00	Martin	696.00	Spain	Male	24.00	9.00	0.00	1.00	0.00	0.00	10883.52	0.00

churn

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00
2	3.00	15619304.00	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	1.00	0.00	113931.57	1.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	0.00	93826.63	0.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
10027	NaN	NaN	Johnstone	516.00	France	Male	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10028	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10029	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10030	9999.00	NaN	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	1.00	0.00	92888.52	1.00
10031	10000.00	15628319.00	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	1.00	0.00	38190.78	0.00

10032 rows × 14 columns

#LETS VISUALIZE THE DISTRIBUTION OF ANY COLUMN JUST TO SEE HOW THE COLUMNS ARE SPREAD
churn['creditscore'].sample(frac=0.001).plot.bar()

<AxesSubplot:>

churn.columns

Index(['creditscore', 'age', 'tenure', 'balance', 'numofproducts',
       'isactivemember', 'estimatedsalary', 'exited', 'geography1',
       'geography_Germany', 'geography_Spain', 'gender_Male'],
      dtype='object')

list(churn.columns)

['creditscore',
 'age',
 'tenure',
 'balance',
 'numofproducts',
 'isactivemember',
 'estimatedsalary',
 'exited',
 'geography1',
 'geography_Germany',
 'geography_Spain',
 'gender_Male']

churn.size

119676

churn.shape

(9973, 12)

print('number of columns',churn.shape[1])

number of columns 12

print('number of rows',churn.shape[0])

number of rows 9973

churn.dtypes

rownumber          float64
customerid         float64
surname             object
creditscore        float64
geography           object
gender              object
age                float64
tenure             float64
balance            float64
numofproducts      float64
hascrcard          float64
isactivemember     float64
estimatedsalary    float64
exited             float64
dtype: object

lets seperate objects from numerical variables

churn_catg=churn.select_dtypes('object')

churn['geography'].value_counts().plot.bar()

<AxesSubplot:>

churn_cont=churn.select_dtypes('number')

churn_catg

	surname	geography	gender
0	Hargrave	France	Female
1	Hill	Spain	Female
2	Onio	France	Female
3	Boni	France	Female
4	NaN	NaN	NaN
...	...	...	...
10027	Johnstone	France	Male
10028	Liu	France	Female
10029	Liu	France	Female
10030	Sabbatini	Germany	Male
10031	Walker	France	Female

10032 rows × 3 columns

churn_cont

	rownumber	customerid	creditscore	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	619.00	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	608.00	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00
2	3.00	15619304.00	502.00	42.00	8.00	159660.80	3.00	1.00	0.00	113931.57	1.00
3	4.00	NaN	699.00	39.00	1.00	0.00	2.00	0.00	0.00	93826.63	0.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...
10027	NaN	NaN	516.00	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10028	NaN	NaN	709.00	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10029	NaN	NaN	709.00	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10030	9999.00	NaN	772.00	42.00	3.00	75075.31	2.00	1.00	0.00	92888.52	1.00
10031	10000.00	15628319.00	792.00	28.00	4.00	130142.79	1.00	1.00	0.00	38190.78	0.00

10032 rows × 11 columns

#UNIQUE VALUES SHOWS THE TYPE OF VALUE IN EACH COLUMN
churn['estimatedsalary'].nunique()

9970

churn['estimatedsalary'].unique

<bound method Series.unique of 0       101348.88
1       112542.58
2       113931.57
3        93826.63
5       149756.71
           ...   
9997     99595.67
10006    96833.00
10025    96270.64
10030    92888.52
10031    38190.78
Name: estimatedsalary, Length: 9973, dtype: float64>

churn['estimatedsalary'].unique()[:20]

array([101348.88, 112542.58, 113931.57,  93826.63,       nan, 149756.71,
       119346.88,  74940.5 ,  71725.73,  80181.12,  76390.01,  26260.98,
       190857.79,  65951.65,  64327.26,   5097.67,  14406.41, 158684.81,
        54724.03, 170886.17])

CLEANING THE DATA

check if the data vave duplicates

churn.duplicated()

0        False
1        False
2        False
3        False
5        False
         ...  
9997     False
10006    False
10025    False
10030    False
10031    False
Length: 9973, dtype: bool

churn.duplicated().any()

True

churn_dup=churn.duplicated().any()

print(churn_dup)

True

churn[churn.duplicated()]

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
7	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
15	14.00	15691483.00	Chin	549.00	France	Female	25.00	5.00	0.00	2.00	0.00	0.00	190857.79	0.00
16	14.00	15691483.00	Chin	549.00	France	Female	25.00	5.00	0.00	2.00	0.00	0.00	190857.79	0.00
17	14.00	15691483.00	Chin	549.00	France	Female	25.00	5.00	0.00	2.00	0.00	0.00	190857.79	0.00
19	15.00	15600882.00	Scott	635.00	Spain	Female	35.00	7.00	0.00	2.00	1.00	1.00	65951.65	0.00
20	15.00	15600882.00	Scott	635.00	Spain	Female	35.00	7.00	0.00	2.00	1.00	1.00	65951.65	0.00
21	15.00	15600882.00	Scott	635.00	Spain	Female	35.00	7.00	0.00	2.00	1.00	1.00	65951.65	0.00
5278	NaN	15607230.00	Michel	588.00	Germany	Male	33.00	9.00	150186.22	2.00	NaN	NaN	NaN	NaN
5280	NaN	15567630.00	Bruce	721.00	Germany	Male	40.00	6.00	100275.88	1.00	NaN	NaN	NaN	NaN
5282	NaN	15587507.00	Feng	850.00	France	Male	47.00	6.00	0.00	1.00	NaN	NaN	NaN	NaN
5284	NaN	15733904.00	McDonald	529.00	France	Male	32.00	9.00	147493.89	1.00	NaN	NaN	NaN	NaN
5286	NaN	NaN	Watt	622.00	France	Male	43.00	8.00	0.00	2.00	NaN	NaN	NaN	NaN
5288	NaN	NaN	Goodwin	683.00	France	Female	42.00	8.00	0.00	2.00	NaN	NaN	NaN	NaN
5290	NaN	NaN	Arcuri	575.00	France	Male	29.00	4.00	121823.40	2.00	NaN	NaN	NaN	NaN
5292	NaN	NaN	Rapuluolisa	577.00	Spain	Female	40.00	1.00	0.00	2.00	NaN	NaN	NaN	NaN
5294	NaN	NaN	Kao	510.00	Germany	Male	31.00	0.00	113688.63	1.00	NaN	NaN	NaN	NaN
5296	NaN	NaN	Tsui	652.00	Spain	Female	40.00	7.00	100471.34	1.00	NaN	NaN	NaN	NaN
9999	NaN	NaN	Burbidge	498.00	Germany	Male	42.00	3.00	152039.70	1.00	NaN	NaN	NaN	NaN
10001	NaN	NaN	Griffin	655.00	Germany	Female	46.00	7.00	137145.12	1.00	NaN	NaN	NaN	NaN
10003	NaN	NaN	Cocci	613.00	France	Male	40.00	4.00	0.00	1.00	NaN	NaN	NaN	NaN
10005	NaN	15696175.00	Echezonachukwu	602.00	Germany	Male	35.00	7.00	90602.42	2.00	NaN	NaN	NaN	NaN
10008	NaN	NaN	Bartlett	673.00	Germany	Male	47.00	1.00	183579.54	2.00	NaN	NaN	NaN	NaN
10010	NaN	NaN	Mancini	606.00	Spain	Male	30.00	8.00	180307.73	2.00	NaN	NaN	NaN	NaN
10012	NaN	NaN	Pirozzi	775.00	France	Male	30.00	4.00	0.00	2.00	NaN	NaN	NaN	NaN
10014	NaN	NaN	McMillan	841.00	Spain	Male	28.00	4.00	0.00	2.00	NaN	NaN	NaN	NaN
10016	NaN	15798964.00	Nkemakonam	714.00	Germany	Male	33.00	3.00	35016.60	1.00	NaN	NaN	NaN	NaN
10018	NaN	NaN	Ajuluchukwu	597.00	France	Female	53.00	4.00	88381.21	1.00	NaN	NaN	NaN	NaN
10020	NaN	NaN	Chukwualuka	726.00	Spain	Male	36.00	2.00	0.00	1.00	NaN	NaN	NaN	NaN
10022	NaN	NaN	Rahman	644.00	France	Male	28.00	7.00	155060.41	1.00	NaN	NaN	NaN	NaN
10024	NaN	NaN	Wood	800.00	France	Female	29.00	2.00	0.00	2.00	NaN	NaN	NaN	NaN
10027	NaN	NaN	Johnstone	516.00	France	Male	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10029	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN

churn.drop_duplicates(inplace=True)

churn_dup=churn.duplicated().any()

print(churn_dup)

False

churn.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 10000 entries, 0 to 10031
Data columns (total 14 columns):
 #   Column           Non-Null Count  Dtype  
---  ------           --------------  -----  
 0   rownumber        9974 non-null   float64
 1   customerid       8515 non-null   float64
 2   surname          9998 non-null   object 
 3   creditscore      9998 non-null   float64
 4   geography        9998 non-null   object 
 5   gender           9998 non-null   object 
 6   age              9998 non-null   float64
 7   tenure           9998 non-null   float64
 8   balance          9998 non-null   float64
 9   numofproducts    9998 non-null   float64
 10  hascrcard        9973 non-null   float64
 11  isactivemember   9973 non-null   float64
 12  estimatedsalary  9973 non-null   float64
 13  exited           9973 non-null   float64
dtypes: float64(11), object(3)
memory usage: 1.1+ MB

churn.count()

rownumber          9974
customerid         8515
surname            9998
creditscore        9998
geography          9998
gender             9998
age                9998
tenure             9998
balance            9998
numofproducts      9998
hascrcard          9973
isactivemember     9973
estimatedsalary    9973
exited             9973
dtype: int64

check for missing values

churn.isnull().sum()

rownumber            26
customerid         1485
surname               2
creditscore           2
geography             2
gender                2
age                   2
tenure                2
balance               2
numofproducts         2
hascrcard            27
isactivemember       27
estimatedsalary      27
exited               27
dtype: int64

churn.isnull().sum().plot.bar()

<AxesSubplot:>

churn.isnull().sum().sum()

1635

churn_missing=churn[churn.isnull().any(1)]

churn_missing

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
1	2.00	NaN	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00
3	4.00	NaN	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	0.00	93826.63	0.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
5	6.00	NaN	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	1.00	0.00	149756.71	1.00
6	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
10023	NaN	NaN	Wood	800.00	France	Female	29.00	2.00	0.00	2.00	NaN	NaN	NaN	NaN
10025	9996.00	NaN	Obijiaku	771.00	France	Male	39.00	5.00	0.00	2.00	1.00	0.00	96270.64	0.00
10026	NaN	NaN	Johnstone	516.00	France	Male	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10028	NaN	NaN	Liu	709.00	France	Female	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10030	9999.00	NaN	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	1.00	0.00	92888.52	1.00

1491 rows × 14 columns

churn.notnull()

	rownumber	customerid	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	True	True	True	True	True	True	True	True	True	True	True	True	True	True
1	True	False	True	True	True	True	True	True	True	True	True	True	True	True
2	True	True	True	True	True	True	True	True	True	True	True	True	True	True
3	True	False	True	True	True	True	True	True	True	True	True	True	True	True
4	True	False	False	False	False	False	False	False	False	False	False	False	False	False
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
10025	True	False	True	True	True	True	True	True	True	True	True	True	True	True
10026	False	False	True	True	True	True	True	True	True	True	False	False	False	False
10028	False	False	True	True	True	True	True	True	True	True	False	False	False	False
10030	True	False	True	True	True	True	True	True	True	True	True	True	True	True
10031	True	True	True	True	True	True	True	True	True	True	True	True	True	True

10000 rows × 14 columns

churn_catg.isnull().sum()

surname      3
geography    3
gender       3
dtype: int64

churn_catg.isnull().sum().plot.bar()

<AxesSubplot:>

churn_cont.isnull().sum()

rownumber            52
customerid         1505
creditscore           3
age                   3
tenure                3
balance               3
numofproducts         3
hascrcard            53
isactivemember       53
estimatedsalary      53
exited               53
dtype: int64

churn_cont.isnull().sum().plot.bar()

<AxesSubplot:>

lets calculate the amount of missing values in percentage of the entire dataset

(len(churn))

10000

(len(churn))*100

1000000

churn_percentage=churn.isnull().sum()/(len(churn))*100

churn_percentage

rownumber          0.26
customerid        14.85
surname            0.02
creditscore        0.02
geography          0.02
gender             0.02
age                0.02
tenure             0.02
balance            0.02
numofproducts      0.02
hascrcard          0.27
isactivemember     0.27
estimatedsalary    0.27
exited             0.27
dtype: float64

(churn_percentage)*round(4)

rownumber          1.04
customerid        59.40
surname            0.08
creditscore        0.08
geography          0.08
gender             0.08
age                0.08
tenure             0.08
balance            0.08
numofproducts      0.08
hascrcard          1.08
isactivemember     1.08
estimatedsalary    1.08
exited             1.08
dtype: float64

VISUALIZING MISSING VALUES

import missingno as msn
msn.matrix(churn)

<AxesSubplot:>

msn.bar(churn)

<AxesSubplot:>

msn.dendrogram(churn)

<AxesSubplot:>

msn.heatmap(churn)

<AxesSubplot:>

churn_nullvalue = pd.DataFrame((churn.isnull().sum())*100/churn.shape[0]).reset_index()
churn_nullvalue.columns = ['Column Name', 'Null Values Percentage']
fig = plt.figure(figsize=(18,6))
ax = sns.pointplot(x="Column Name",y="Null Values Percentage",data=churn_nullvalue,color='green')
plt.xticks(rotation =90,fontsize =8)
ax.axhline(40, ls='--',color='red')
plt.title("Percentage of Missing values in application data")
plt.ylabel("Null Values PERCENTAGE")
plt.xlabel("COLUMNS")
plt.show()

sns.heatmap(churn.isnull())

<AxesSubplot:>

churn.isnull().sum()

rownumber            26
customerid         1485
surname               2
creditscore           2
geography             2
gender                2
age                   2
tenure                2
balance               2
numofproducts         2
hascrcard            27
isactivemember       27
estimatedsalary      27
exited               27
dtype: int64

churn=churn.drop(labels=['customerid'],axis=1)

churn.isnull().sum()

rownumber          26
surname             2
creditscore         2
geography           2
gender              2
age                 2
tenure              2
balance             2
numofproducts       2
hascrcard          27
isactivemember     27
estimatedsalary    27
exited             27
dtype: int64

churn=churn.drop(labels=['rownumber','hascrcard'],axis=1)

churn.isnull().sum()

surname             2
creditscore         2
geography           2
gender              2
age                 2
tenure              2
balance             2
numofproducts       2
isactivemember     27
estimatedsalary    27
exited             27
dtype: int64

churn.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

len(churn.columns)

11

churn_cont.isnull().sum().plot.bar()

<AxesSubplot:>

INPUTATION

churn_cont

	rownumber	customerid	creditscore	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	619.00	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	608.00	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00
2	3.00	15619304.00	502.00	42.00	8.00	159660.80	3.00	1.00	0.00	113931.57	1.00
3	4.00	NaN	699.00	39.00	1.00	0.00	2.00	0.00	0.00	93826.63	0.00
4	5.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...
10027	NaN	NaN	516.00	35.00	10.00	57369.61	1.00	NaN	NaN	NaN	NaN
10028	NaN	NaN	709.00	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10029	NaN	NaN	709.00	36.00	7.00	0.00	1.00	NaN	NaN	NaN	NaN
10030	9999.00	NaN	772.00	42.00	3.00	75075.31	2.00	1.00	0.00	92888.52	1.00
10031	10000.00	15628319.00	792.00	28.00	4.00	130142.79	1.00	1.00	0.00	38190.78	0.00

10032 rows × 11 columns

churn.isnull().sum()

surname             2
creditscore         2
geography           2
gender              2
age                 2
tenure              2
balance             2
numofproducts       2
isactivemember     27
estimatedsalary    27
exited             27
dtype: int64

churn.isnull().sum().sum()

97

mean_churn1=churn['estimatedsalary'].mean()

mean_churn1

100093.52722149782

churn['estimatedsalary'].fillna(mean_churn1,inplace=True)

churn.isnull().sum()

surname             2
creditscore         2
geography           2
gender              2
age                 2
tenure              2
balance             2
numofproducts       2
isactivemember     27
estimatedsalary     0
exited             27
dtype: int64

churn.isnull().sum().sum()

70

churn['isactivemember'].unique()

array([ 1.,  0., nan])

churn.isnull().sum()

surname             2
creditscore         2
geography           2
gender              2
age                 2
tenure              2
balance             2
numofproducts       2
isactivemember     27
estimatedsalary     0
exited             27
dtype: int64

churn.isnull().sum().sum()

70

churn_cont.head(2)

	rownumber	customerid	creditscore	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	619.00	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	608.00	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00

#drop all categorical or object or string missing values
churn.loc[:,['surname','geography',]].dropna()

	surname	geography
0	Hargrave	France
1	Hill	Spain
2	Onio	France
3	Boni	France
5	Chu	Spain
...	...	...
10025	Obijiaku	France
10026	Johnstone	France
10028	Liu	France
10030	Sabbatini	Germany
10031	Walker	France

9998 rows × 2 columns

churn.isnull().sum()

surname             2
creditscore         2
geography           2
gender              2
age                 2
tenure              2
balance             2
numofproducts       2
isactivemember     27
estimatedsalary     0
exited             27
dtype: int64

churn.gender.unique()

array(['Female', nan, 'Male'], dtype=object)

len(churn.gender.unique())

3

churn_mod=churn.gender.value_counts()

churn_mod

Male      5456
Female    4542
Name: gender, dtype: int64

#mode is male
churn['gender'].fillna('Male',inplace=True)

churn['gender'].fillna('churn_mod',inplace=True)

churn.isnull().sum()

surname             2
creditscore         2
geography           2
gender              0
age                 2
tenure              2
balance             2
numofproducts       2
isactivemember     27
estimatedsalary     0
exited             27
dtype: int64

churn.isnull().sum().sum()

68

churn.dropna(inplace=True)

churn.isnull().sum()

surname            0
creditscore        0
geography          0
gender             0
age                0
tenure             0
balance            0
numofproducts      0
isactivemember     0
estimatedsalary    0
exited             0
dtype: int64

churn.isnull().sum().sum()

0

VISUALIZING THE DATA AFTER TAKING CARE OF THE MISSING VALUES

#check new visuals
msn.matrix(churn)

<AxesSubplot:>

msn.bar(churn)

<AxesSubplot:>

msn.dendrogram(churn)

<AxesSubplot:>

msn.heatmap(churn)

<AxesSubplot:>

churn_nullvalue = pd.DataFrame((churn.isnull().sum())*100/churn.shape[0]).reset_index()
churn_nullvalue.columns = ['Column Name', 'Null Values Percentage']
fig = plt.figure(figsize=(18,6))
ax = sns.pointplot(x="Column Name",y="Null Values Percentage",data=churn_nullvalue,color='green')
plt.xticks(rotation =90,fontsize =8)
ax.axhline(40, ls='--',color='red')
plt.title("Percentage of Missing values in application data")
plt.ylabel("Null Values PERCENTAGE")
plt.xlabel("COLUMNS")
plt.show()

sns.heatmap(churn.isnull())

<AxesSubplot:>

OUTLIERS

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn.describe()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
count	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00
mean	650.49	38.93	5.01	76498.54	1.53	0.52	100093.53	0.20
std	96.63	10.50	2.89	62383.80	0.58	0.50	57494.40	0.40
min	350.00	18.00	0.00	0.00	1.00	0.00	11.58	0.00
25%	584.00	32.00	3.00	0.00	1.00	0.00	51099.56	0.00
50%	652.00	37.00	5.00	97234.58	1.00	1.00	100187.43	0.00
75%	717.00	44.00	7.00	127642.44	2.00	1.00	149381.32	0.00
max	850.00	92.00	10.00	250898.09	4.00	1.00	199992.48	1.00

sns.distplot(churn['balance'],kde=False,hist=True,bins=12)
plt.title('balance distribution',size=16)
plt.ylabel('count')

Text(0, 0.5, 'count')

churn.boxplot()

<AxesSubplot:>

sns.distplot(churn['balance'])

<AxesSubplot:xlabel='balance', ylabel='Density'>

churn['estimatedsalary'].plot(kind='box',figsize=(3,4),patch_artist=True)

<AxesSubplot:>

sns.boxplot(x='gender',y='age',data=churn)

<AxesSubplot:xlabel='gender', ylabel='age'>

DETECT OUTLIERS

Z-SCORE METHOD

churn['balance'].plot(kind='box',figsize=(3,4),patch_artist=True)

<AxesSubplot:>

churn['age'].unique()

array([42., 41., 39., 44., 29., 27., 31., 24., 34., 25., 35., 45., 58.,
       32., 38., 46., 36., 43., 33., 40., 51., 61., 49., 37., 19., 66.,
       56., 26., 21., 55., 75., 22., 30., 28., 65., 48., 52., 50., 57.,
       73., 47., 54., 72., 20., 67., 79., 62., 53., 80., 59., 68., 23.,
       60., 70., 63., 64., 18., 82., 69., 74., 71., 76., 77., 88., 85.,
       84., 78., 81., 92., 83.])

sns.boxplot(y='age',data=churn)

<AxesSubplot:ylabel='age'>

churn['age'].mean()

38.9250977639627

churn1=churn[churn['age']<=70]

sns.boxplot(y='age',data=churn1)

<AxesSubplot:ylabel='age'>

churn2=churn[churn['age']<=58]

sns.boxplot(y='age',data=churn2)

<AxesSubplot:ylabel='age'>

sns.boxplot(y='balance',data=churn)

<AxesSubplot:ylabel='balance'>

churn2=churn

churn['balance'].mean()

76498.54009425409

churn.shape

(9973, 11)

churn.boxplot()

<AxesSubplot:>

churn.hist()

array([[<AxesSubplot:title={'center':'creditscore'}>,
        <AxesSubplot:title={'center':'age'}>,
        <AxesSubplot:title={'center':'tenure'}>],
       [<AxesSubplot:title={'center':'balance'}>,
        <AxesSubplot:title={'center':'numofproducts'}>,
        <AxesSubplot:title={'center':'isactivemember'}>],
       [<AxesSubplot:title={'center':'estimatedsalary'}>,
        <AxesSubplot:title={'center':'exited'}>, <AxesSubplot:>]],
      dtype=object)

DETECT AND REMOVE OUTLIERS

Data points, that falls outside of 1.5 times of the interquartile range above the 3rd or below the 1st quartile

is an outlier.

churn.describe()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
count	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00
mean	650.49	38.93	5.01	76498.54	1.53	0.52	100093.53	0.20
std	96.63	10.50	2.89	62383.80	0.58	0.50	57494.40	0.40
min	350.00	18.00	0.00	0.00	1.00	0.00	11.58	0.00
25%	584.00	32.00	3.00	0.00	1.00	0.00	51099.56	0.00
50%	652.00	37.00	5.00	97234.58	1.00	1.00	100187.43	0.00
75%	717.00	44.00	7.00	127642.44	2.00	1.00	149381.32	0.00
max	850.00	92.00	10.00	250898.09	4.00	1.00	199992.48	1.00

churn.shape

(9973, 11)

BOXPLOT

#DEFINE A FUNCTION CALLED plot_box 
#and there will be two arguments call it df and ft
#and plot the boxplot with the boxplot function in pandas
def plot_boxplot(churn1,ft):
    churn.boxplot(column=[ft])
    plt.grid(False)
    plt.show()

#lets call it
plot_boxplot(churn,'creditscore')

sns.boxplot(y='age',data=churn1)

<AxesSubplot:ylabel='age'>

sns.boxplot(y='creditscore',data=churn1)

<AxesSubplot:ylabel='creditscore'>

plot_boxplot(churn,'tenure')

plot_boxplot(churn,'balance')

plot_boxplot(churn,'numofproducts')

churn1=churn[churn['numofproducts']<=3.5]

plot_boxplot(churn1,'numofproducts')

plot_boxplot(churn,'estimatedsalary')

REMOVE OUTLIERS

#FIRST WE WILL EXTRAT ALL OUTLIERS IN ALL COLUMN 
# SECONDLY WE USE THE INDEXES OF THE OUTLIERS TO REMOVE THEM FROM THE DATAFRAME  BY CREATING A LIST 
#DEFINE A FUNCTION CALLED OULIERS
# IQR =Q3 -  Q1 
# +/- 1.5 * IQR
# SET THE UPPER AND LOWER BOUNDRIES

Q1=churn.creditscore.quantile(0.25)
Q3=churn.creditscore.quantile(0.75)

Q1,Q3

(584.0, 717.0)

plt.hist(churn['creditscore'],)
plt.title('distribution of total creditscore')
plt.show()

sns.distplot(churn['creditscore'])

<AxesSubplot:xlabel='creditscore', ylabel='Density'>

IQR=Q3-Q1 

IQR

133.0

lower_limit = Q1-(1.5*IQR)

lower_limit

384.5

upper_limit=Q3+(1.5*IQR)

upper_limit

916.5

lower_limit, upper_limit

(384.5, 916.5)

churn[(churn.creditscore<lower_limit)|(churn.creditscore>upper_limit)]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
8	Obinna	376.00	Germany	Female	29.00	4.00	115046.74	4.00	0.00	119346.88	1.00
949	Lin	376.00	France	Female	46.00	6.00	0.00	1.00	0.00	157333.69	1.00
1200	Thomas	363.00	Spain	Female	28.00	6.00	146098.43	3.00	0.00	100615.14	1.00
1412	Panicucci	359.00	France	Female	44.00	6.00	128747.69	1.00	0.00	146955.71	1.00
1638	Azubuike	350.00	Spain	Male	54.00	1.00	152677.48	1.00	1.00	191973.49	1.00
1845	Campbell	350.00	Germany	Male	39.00	0.00	109733.20	2.00	0.00	123602.11	1.00
1969	Aikenhead	358.00	Spain	Female	52.00	8.00	143542.36	3.00	0.00	141959.11	1.00
2480	Chou	351.00	Germany	Female	57.00	4.00	163146.46	1.00	0.00	169621.69	1.00
2586	Ozoemena	365.00	Germany	Male	30.00	0.00	127760.07	1.00	0.00	81537.85	1.00
5511	Ijendu	383.00	Spain	Female	48.00	8.00	95808.19	1.00	0.00	137702.01	1.00
8171	Ch'ien	367.00	Spain	Male	42.00	6.00	93608.28	1.00	0.00	168816.73	1.00
8740	Onyekachi	350.00	France	Male	51.00	10.00	0.00	1.00	1.00	125823.79	1.00
8779	Lin	350.00	France	Female	60.00	3.00	0.00	1.00	0.00	113796.15	1.00
9227	Watts	382.00	Spain	Male	36.00	0.00	0.00	1.00	1.00	179540.73	1.00
9373	Loggia	373.00	France	Male	42.00	7.00	0.00	1.00	0.00	77786.37	1.00
9641	Maslow	350.00	France	Female	40.00	0.00	111098.85	1.00	1.00	172321.21	1.00

(churn['creditscore']<916.5)

0        True
1        True
2        True
3        True
5        True
         ... 
9997     True
10006    True
10025    True
10030    True
10031    True
Name: creditscore, Length: 9973, dtype: bool

(churn['creditscore']>384.5)

0        True
1        True
2        True
3        True
5        True
         ... 
9997     True
10006    True
10025    True
10030    True
10031    True
Name: creditscore, Length: 9973, dtype: bool

churn2=churn[(churn['creditscore']<916.5)&(churn['creditscore']>384.5)]

churn2['creditscore'].plot(kind='box', figsize=(3,4), patch_artist=True)

<AxesSubplot:>

churn2.shape

(9957, 11)

plt.hist(churn2['creditscore'])
plt.title('distribution of total creditscore')
plt.show()

sns.distplot(churn2['creditscore'])

<AxesSubplot:xlabel='creditscore', ylabel='Density'>

churn2=churn

#cross check by visualizing the box plot
plot_boxplot(churn,'creditscore')

plot_boxplot(churn,'age')

#SAVE THE NEW DATA FRAME IN PANDAS
churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn.shape

(9973, 11)

churn.describe()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
count	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00	9973.00
mean	650.49	38.93	5.01	76498.54	1.53	0.52	100093.53	0.20
std	96.63	10.50	2.89	62383.80	0.58	0.50	57494.40	0.40
min	350.00	18.00	0.00	0.00	1.00	0.00	11.58	0.00
25%	584.00	32.00	3.00	0.00	1.00	0.00	51099.56	0.00
50%	652.00	37.00	5.00	97234.58	1.00	1.00	100187.43	0.00
75%	717.00	44.00	7.00	127642.44	2.00	1.00	149381.32	0.00
max	850.00	92.00	10.00	250898.09	4.00	1.00	199992.48	1.00

churn.boxplot()

<AxesSubplot:>

sns.pairplot(churn)

<seaborn.axisgrid.PairGrid at 0x1b564768b20>

EXPLORATORY DATA ANALYSIS

churn.corr()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
creditscore	1.00	-0.00	0.00	0.01	0.01	0.03	-0.00	-0.03
age	-0.00	1.00	-0.01	0.03	-0.03	0.09	-0.01	0.29
tenure	0.00	-0.01	1.00	-0.01	0.01	-0.03	0.01	-0.01
balance	0.01	0.03	-0.01	1.00	-0.30	-0.01	0.01	0.12
numofproducts	0.01	-0.03	0.01	-0.30	1.00	0.01	0.01	-0.05
isactivemember	0.03	0.09	-0.03	-0.01	0.01	1.00	-0.01	-0.16
estimatedsalary	-0.00	-0.01	0.01	0.01	0.01	-0.01	1.00	0.01
exited	-0.03	0.29	-0.01	0.12	-0.05	-0.16	0.01	1.00

plt.figure(figsize=(20,20))
sns.heatmap(churn.corr(), vmin=-1, cmap="plasma_r", annot=True)
#same thing can be seen from the correlation as well

<AxesSubplot:>

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

form the corrolation heatmap it shows exited which is the target variable is highly corrolated with the following independent variable, estimated salary , balance and age.

EXPLORATORY DATA ANALYSIS

QUERRYING THE DATASET

Q1,WHAT IS THE MEAN, MIN , AND MAX VALUE OF estimatedsalary

GROUPBY

churn.groupby('estimatedsalary').mean()

	creditscore	age	tenure	balance	numofproducts	isactivemember	exited
estimatedsalary
11.58	709.00	45.00	4.00	122917.71	1.00	1.00	1.00
90.07	625.00	31.00	5.00	0.00	2.00	1.00	0.00
91.75	645.00	59.00	8.00	121669.93	2.00	0.00	1.00
96.27	710.00	38.00	2.00	0.00	2.00	0.00	0.00
106.67	629.00	40.00	9.00	0.00	1.00	0.00	0.00
...	...	...	...	...	...	...	...
199909.32	620.00	28.00	8.00	0.00	2.00	1.00	0.00
199929.17	569.00	37.00	9.00	178755.84	1.00	0.00	0.00
199953.33	608.00	27.00	4.00	153325.10	1.00	1.00	0.00
199970.74	639.00	41.00	5.00	98635.77	1.00	0.00	0.00
199992.48	685.00	42.00	2.00	0.00	2.00	0.00	0.00

9970 rows × 7 columns

churn.groupby('creditscore').min()

	surname	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
creditscore
350.00	Azubuike	France	Female	39.00	0.00	0.00	1.00	0.00	113796.15	1.00
351.00	Chou	Germany	Female	57.00	4.00	163146.46	1.00	0.00	169621.69	1.00
358.00	Aikenhead	Spain	Female	52.00	8.00	143542.36	3.00	0.00	141959.11	1.00
359.00	Panicucci	France	Female	44.00	6.00	128747.69	1.00	0.00	146955.71	1.00
363.00	Thomas	Spain	Female	28.00	6.00	146098.43	3.00	0.00	100615.14	1.00
...	...	...	...	...	...	...	...	...	...	...
846.00	Baxter	France	Female	34.00	0.00	0.00	1.00	0.00	3440.47	0.00
847.00	Beam	France	Female	31.00	3.00	0.00	1.00	0.00	16025.17	0.00
848.00	Hung	Germany	Female	22.00	3.00	90018.45	1.00	0.00	30876.84	0.00
849.00	Chou	France	Female	27.00	0.00	0.00	1.00	0.00	17294.12	0.00
850.00	Abbott	France	Female	19.00	0.00	0.00	1.00	0.00	705.18	0.00

460 rows × 10 columns

Q2, FIND ALL INSTANCES WHEN CREDITSCORE IS GREATER THAN OR EQUALL TO 600 OR GEOGRAPHY IS FRANCE

churn[(churn['creditscore']>=600)|(churn['geography']=='france')]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
3	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
10	H?	684.00	France	Male	27.00	2.00	134603.88	1.00	1.00	71725.73	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00
10006	Nepean	659.00	France	Male	36.00	6.00	123841.49	2.00	0.00	96833.00	0.00
10025	Obijiaku	771.00	France	Male	39.00	5.00	0.00	2.00	0.00	96270.64	0.00
10030	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	0.00	92888.52	1.00
10031	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	0.00	38190.78	0.00

6947 rows × 11 columns

list[churn[(churn['creditscore']<600)|(churn['geography']=='france')]]

list[       surname  creditscore geography  gender   age  tenure   balance  \
2         Onio       502.00    France  Female 42.00    8.00 159660.80   
8       Obinna       376.00   Germany  Female 29.00    4.00 115046.74   
9           He       501.00    France    Male 44.00    4.00 142051.07   
11      Bearce       528.00    France    Male 31.00    6.00 102016.72   
12     Andrews       497.00     Spain    Male 24.00    3.00      0.00   
...        ...          ...       ...     ...   ...     ...       ...   
9981   Douglas       479.00    France    Male 34.00    5.00 117593.48   
9983     Ch'en       479.00     Spain    Male 35.00    4.00 125920.98   
9985  McGregor       512.00   Germany  Female 40.00    8.00 153537.57   
9987  Thompson       518.00    France    Male 42.00    7.00 151027.05   
9991    Barker       583.00    France    Male 33.00    7.00 122531.86   

      numofproducts  isactivemember  estimatedsalary  exited  
2              3.00            0.00        113931.57    1.00  
8              4.00            0.00        119346.88    1.00  
9              2.00            1.00         74940.50    0.00  
11             2.00            0.00         80181.12    0.00  
12             2.00            0.00         76390.01    0.00  
...             ...             ...              ...     ...  
9981           2.00            0.00        113308.29    0.00  
9983           1.00            1.00         20393.44    0.00  
9985           2.00            0.00         23101.13    0.00  
9987           2.00            0.00        119377.36    0.00  
9991           1.00            0.00         13549.24    0.00  

[3026 rows x 11 columns]]

Q3 SHOW ALL RECORDS WHERE CREDITSTORE IS 502.0 OR 518.0

IS IN

churn['creditscore'].isin(['502.0','518.0'])

0        False
1        False
2        False
3        False
5        False
         ...  
9997     False
10006    False
10025    False
10030    False
10031    False
Name: creditscore, Length: 9973, dtype: bool

QUESTION 4 FIND TOTAL NUMBER OF RECORDS HAVING CREDITSCORE OF 502 TO 518 INCLUSIVE USING BETWEEN METHOD

[churn['creditscore']>=502]

[0        True
 1        True
 2        True
 3        True
 5        True
          ... 
 9997     True
 10006    True
 10025    True
 10030    True
 10031    True
 Name: creditscore, Length: 9973, dtype: bool]

[churn['creditscore']<=518]

[0        False
 1        False
 2         True
 3        False
 5        False
          ...  
 9997     False
 10006    False
 10025    False
 10030    False
 10031    False
 Name: creditscore, Length: 9973, dtype: bool]

sum(churn['creditscore'].between(502,518))

271

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

geography

Q5 how many people have spain as thier geographical location

churn['geography']

0         France
1          Spain
2         France
3         France
5          Spain
          ...   
9997       Spain
10006     France
10025     France
10030    Germany
10031     France
Name: geography, Length: 9973, dtype: object

churn['geography'].str.contains('spain',case=False)

0        False
1         True
2        False
3        False
5         True
         ...  
9997      True
10006    False
10025    False
10030    False
10031    False
Name: geography, Length: 9973, dtype: bool

churn[churn['geography'].str.contains('spain',case=False)]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
12	Andrews	497.00	Spain	Male	24.00	3.00	0.00	2.00	0.00	76390.01	0.00
18	Scott	635.00	Spain	Female	35.00	7.00	0.00	2.00	1.00	65951.65	0.00
24	Henderson	549.00	Spain	Female	24.00	9.00	0.00	2.00	1.00	14406.41	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9976	Graham	764.00	Spain	Male	38.00	4.00	113607.47	1.00	0.00	91094.46	0.00
9978	Aldridge	573.00	Spain	Male	44.00	9.00	0.00	2.00	0.00	107124.17	0.00
9979	Flynn	702.00	Spain	Male	44.00	9.00	0.00	1.00	0.00	59207.41	1.00
9983	Ch'en	479.00	Spain	Male	35.00	4.00	125920.98	1.00	1.00	20393.44	0.00
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00

2471 rows × 11 columns

len(churn['geography'].str.contains('spain',case=False))

9973

Q6 TOTAL NUMBER OF RECORDS IN GEOGRAPHY

churn['geography'].nunique()

3

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

Q7 WHICH GEOGRAPHY HAS THE HIGEST CREDITSCORE AND WHICH GEOGRAPHY HAVE THE HIGHEST AVERAGE RATING

churn.groupby('geography')['creditscore'].max()

geography
France    850.00
Germany   850.00
Spain     850.00
Name: creditscore, dtype: float64

churn.groupby('geography')['creditscore'].mean()

geography
France    649.61
Germany   651.55
Spain     651.19
Name: creditscore, dtype: float64

churn.groupby('geography')['creditscore'].mean().sort_values(ascending=False)

geography
Germany   651.55
Spain     651.19
France    649.61
Name: creditscore, dtype: float64

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

Q8 TOTAL NUMBER OF GEOGRAPHY HAVING 2 TENURE

churn.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

churn['tenure']

0       2.00
1       1.00
2       8.00
3       1.00
5       8.00
        ... 
9997    6.00
10006   6.00
10025   5.00
10030   3.00
10031   4.00
Name: tenure, Length: 9973, dtype: float64

churn['tenure']==2

0         True
1        False
2        False
3        False
5        False
         ...  
9997     False
10006    False
10025    False
10030    False
10031    False
Name: tenure, Length: 9973, dtype: bool

churn[churn['tenure']==2]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
10	H?	684.00	France	Male	27.00	2.00	134603.88	1.00	1.00	71725.73	0.00
33	Young	756.00	Germany	Male	36.00	2.00	136815.64	1.00	1.00	170041.95	0.00
49	Osborne	556.00	France	Female	61.00	2.00	117419.35	1.00	1.00	94153.83	0.00
50	Lavine	834.00	France	Female	49.00	2.00	131394.56	1.00	0.00	194365.76	1.00
...	...	...	...	...	...	...	...	...	...	...	...
9922	Emenike	645.00	Germany	Male	41.00	2.00	93925.30	1.00	0.00	123982.14	1.00
9926	Endrizzi	739.00	France	Male	58.00	2.00	101579.28	1.00	1.00	72168.53	0.00
9960	Fu	784.00	Spain	Male	23.00	2.00	0.00	1.00	1.00	6847.73	0.00
9970	Burke	550.00	France	Male	47.00	2.00	0.00	2.00	1.00	97057.28	0.00
9980	Scott	740.00	Germany	Male	33.00	2.00	126524.11	1.00	0.00	136869.31	0.00

1045 rows × 11 columns

len(churn[churn['tenure']==2])

1045

GENDER

Q9 FIND TOTAL NUMBER OF MALE AND FEMALE

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn['gender'].value_counts()

Male      5438
Female    4535
Name: gender, dtype: int64

Q10 WHICH GENDER HAVE THE HIGEST BALANCE

churn['balance'].max()

250898.09

churn[churn['balance'].max()==churn['balance']]['gender']

2099    Male
Name: gender, dtype: object

QUESTION11, DISPLAY TOP 5 NAMES HAVING MAXIMUM NUMBER OF SALARY

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn_index=churn['estimatedsalary'].sort_values(ascending=False).head().index

churn.iloc[churn_index]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
6692	Christmas	654.00	Spain	Female	36.00	5.00	0.00	2.00	0.00	157238.05	0.00
3596	Watts	446.00	France	Female	51.00	4.00	105056.13	1.00	0.00	70613.52	0.00
4433	Chao	707.00	France	Female	44.00	6.00	0.00	2.00	1.00	192542.17	0.00
9142	Iheanacho	749.00	Spain	Male	34.00	2.00	0.00	1.00	0.00	174189.04	1.00
8972	Afanasyeva	518.00	Spain	Male	50.00	4.00	0.00	1.00	0.00	107112.25	1.00

QUESTION12 WHAT IS THE RECORDS OF THE SURNAME WATTS

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn['surname']

0         Hargrave
1             Hill
2             Onio
3             Boni
5              Chu
           ...    
9997          T'ao
10006       Nepean
10025     Obijiaku
10030    Sabbatini
10031       Walker
Name: surname, Length: 9973, dtype: object

churn['surname'].isin(['Watts'])

0        False
1        False
2        False
3        False
5        False
         ...  
9997     False
10006    False
10025    False
10030    False
10031    False
Name: surname, Length: 9973, dtype: bool

churn[churn['surname'].isin(['Watts'])]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
529	Watts	527.00	Spain	Male	37.00	5.00	93722.73	2.00	1.00	139093.73	0.00
3596	Watts	446.00	France	Female	51.00	4.00	105056.13	1.00	0.00	70613.52	0.00
9227	Watts	382.00	Spain	Male	36.00	0.00	0.00	1.00	1.00	179540.73	1.00
9635	Watts	581.00	France	Female	24.00	3.00	95508.20	1.00	1.00	45755.00	0.00
9826	Watts	820.00	France	Female	39.00	1.00	104614.29	1.00	0.00	61538.43	1.00

# WE CAN ALSO USE STR CONTAINS
churn[churn['surname'].str.contains('Watts')]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
529	Watts	527.00	Spain	Male	37.00	5.00	93722.73	2.00	1.00	139093.73	0.00
3596	Watts	446.00	France	Female	51.00	4.00	105056.13	1.00	0.00	70613.52	0.00
9227	Watts	382.00	Spain	Male	36.00	0.00	0.00	1.00	1.00	179540.73	1.00
9635	Watts	581.00	France	Female	24.00	3.00	95508.20	1.00	1.00	45755.00	0.00
9826	Watts	820.00	France	Female	39.00	1.00	104614.29	1.00	0.00	61538.43	1.00

AGE

Q13 DISPLAY TOP 5 OLDEST GENDER

churn['age'].unique

<bound method Series.unique of 0       42.00
1       41.00
2       42.00
3       39.00
5       44.00
         ... 
9997    35.00
10006   36.00
10025   39.00
10030   42.00
10031   28.00
Name: age, Length: 9973, dtype: float64>

churn['age'].sort_values(ascending=False)

6776   92.00
6460   92.00
2465   88.00
3040   85.00
3538   84.00
        ... 
9949   18.00
2143   18.00
2148   18.00
4563   18.00
2028   18.00
Name: age, Length: 9973, dtype: float64

churn['age'].sort_values(ascending=False).head()

6776   92.00
6460   92.00
2465   88.00
3040   85.00
3538   84.00
Name: age, dtype: float64

index=churn['age'].sort_values(ascending=False).head().index

churn.iloc[index]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
6805	Costa	825.00	Germany	Male	37.00	6.00	118050.79	1.00	1.00	52301.15	0.00
6489	Johnston	651.00	Germany	Female	34.00	10.00	148962.46	1.00	0.00	66389.43	1.00
2474	Yuryeva	617.00	France	Female	27.00	4.00	0.00	2.00	0.00	190269.21	0.00
3049	Mahon	835.00	Germany	Female	29.00	10.00	130420.20	2.00	0.00	106276.55	0.00
3547	Crawford	672.00	France	Male	43.00	5.00	0.00	1.00	0.00	63833.09	0.00

Q14 GIVE ALL THE NAMES THAT HAVE AGE 50

churn['age']

0       42.00
1       41.00
2       42.00
3       39.00
5       44.00
         ... 
9997    35.00
10006   36.00
10025   39.00
10030   42.00
10031   28.00
Name: age, Length: 9973, dtype: float64

churn[churn['age']==50.0]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
142	Arthur	589.00	Germany	Female	50.00	5.00	144895.05	2.00	1.00	34941.23	0.00
191	Cocci	804.00	France	Female	50.00	4.00	0.00	1.00	1.00	8546.87	1.00
209	He	516.00	Spain	Male	50.00	5.00	0.00	1.00	1.00	146145.93	1.00
235	O'Sullivan	751.00	Germany	Male	50.00	2.00	96888.39	1.00	0.00	77206.25	1.00
315	Balashov	614.00	France	Male	50.00	4.00	137104.47	1.00	0.00	127166.49	1.00
...	...	...	...	...	...	...	...	...	...	...	...
9362	Miah	582.00	France	Male	50.00	2.00	148942.00	1.00	1.00	116944.30	0.00
9655	Baldwin	744.00	Germany	Female	50.00	1.00	121498.11	2.00	1.00	106061.47	1.00
9899	Okeke	771.00	France	Male	50.00	3.00	105229.72	1.00	1.00	16281.68	1.00
9941	Palerma	632.00	Germany	Female	50.00	5.00	107959.39	1.00	1.00	6985.34	1.00
9992	Smith	610.00	Germany	Male	50.00	1.00	113957.01	2.00	0.00	196526.55	1.00

133 rows × 11 columns

len(churn[churn['age']==50.0])

133

Q15 SHOW ALL THE RECORDS THAT HAVE BALANCE ABOVE 83807.86

BALANCE

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn['balance']>=83807.86

0        False
1         True
2         True
3        False
5         True
         ...  
9997     False
10006     True
10025    False
10030    False
10031     True
Name: balance, Length: 9973, dtype: bool

churn[churn['balance']>=83807.86]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
8	Obinna	376.00	Germany	Female	29.00	4.00	115046.74	4.00	0.00	119346.88	1.00
9	He	501.00	France	Male	44.00	4.00	142051.07	2.00	1.00	74940.50	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9993	Azikiwe	637.00	France	Female	33.00	7.00	103377.81	1.00	0.00	84419.78	0.00
9995	P'eng	774.00	France	Male	40.00	9.00	93017.47	2.00	0.00	191608.97	0.00
9996	Diribe	677.00	France	Female	58.00	1.00	90022.85	1.00	1.00	2988.28	0.00
10006	Nepean	659.00	France	Male	36.00	6.00	123841.49	2.00	0.00	96833.00	0.00
10031	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	0.00	38190.78	0.00

5627 rows × 11 columns

churn['balance'].unique()

array([     0.  ,  83807.86, 159660.8 , ..., 123841.49,  75075.31,
       130142.79])

QUESTION 16 SHOW ONLY THE BALANCE OF ALL GENDER FEMALE WHO LIVES IN FRANCE ONLY

(churn['gender']=='Female')

0         True
1         True
2         True
3         True
5        False
         ...  
9997     False
10006    False
10025    False
10030    False
10031     True
Name: gender, Length: 9973, dtype: bool

(churn['geography']=='France')

0         True
1        False
2         True
3         True
5        False
         ...  
9997     False
10006     True
10025     True
10030    False
10031     True
Name: geography, Length: 9973, dtype: bool

(churn['gender']=='Female')&(churn['geography']=='France')

0         True
1        False
2         True
3         True
5        False
         ...  
9997     False
10006    False
10025    False
10030    False
10031     True
Length: 9973, dtype: bool

churn[(churn['gender']=='Female')&(churn['geography']=='France')]['balance']

0            0.00
2       159660.80
3            0.00
13           0.00
14           0.00
           ...   
9988    144751.81
9993    103377.81
9994         0.00
9996     90022.85
10031   130142.79
Name: balance, Length: 2257, dtype: float64

QUESTION 17 SHOW ALL RECORDS WERE GENDER IS MALE AND GEOGRAPHY IS SPAIN OR BALANCE IS 159660.8

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn['gender'].unique()

array(['Female', 'Male'], dtype=object)

(churn['gender']=='Male')

0        False
1        False
2        False
3        False
5         True
         ...  
9997      True
10006     True
10025     True
10030     True
10031    False
Name: gender, Length: 9973, dtype: bool

(churn['geography']=='Spain')

0        False
1         True
2        False
3        False
5         True
         ...  
9997      True
10006    False
10025    False
10030    False
10031    False
Name: geography, Length: 9973, dtype: bool

(churn['gender']=='Male')&(churn['geography']=='Spain')

0        False
1        False
2        False
3        False
5         True
         ...  
9997      True
10006    False
10025    False
10030    False
10031    False
Length: 9973, dtype: bool

churn[(churn['gender']=='Male')&(churn['geography']=='Spain')]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
12	Andrews	497.00	Spain	Male	24.00	3.00	0.00	2.00	0.00	76390.01	0.00
25	Muldrow	587.00	Spain	Male	45.00	6.00	0.00	1.00	0.00	158684.81	0.00
43	Watson	490.00	Spain	Male	31.00	3.00	145260.23	1.00	1.00	114066.77	0.00
44	Lorenzo	804.00	Spain	Male	33.00	7.00	76548.60	1.00	1.00	98453.45	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9976	Graham	764.00	Spain	Male	38.00	4.00	113607.47	1.00	0.00	91094.46	0.00
9978	Aldridge	573.00	Spain	Male	44.00	9.00	0.00	2.00	0.00	107124.17	0.00
9979	Flynn	702.00	Spain	Male	44.00	9.00	0.00	1.00	0.00	59207.41	1.00
9983	Ch'en	479.00	Spain	Male	35.00	4.00	125920.98	1.00	1.00	20393.44	0.00
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00

1385 rows × 11 columns

churn[(churn['gender']=='Male')&(churn['geography']=='Spain')|(churn['balance']=='159660.8')]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
12	Andrews	497.00	Spain	Male	24.00	3.00	0.00	2.00	0.00	76390.01	0.00
25	Muldrow	587.00	Spain	Male	45.00	6.00	0.00	1.00	0.00	158684.81	0.00
43	Watson	490.00	Spain	Male	31.00	3.00	145260.23	1.00	1.00	114066.77	0.00
44	Lorenzo	804.00	Spain	Male	33.00	7.00	76548.60	1.00	1.00	98453.45	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9976	Graham	764.00	Spain	Male	38.00	4.00	113607.47	1.00	0.00	91094.46	0.00
9978	Aldridge	573.00	Spain	Male	44.00	9.00	0.00	2.00	0.00	107124.17	0.00
9979	Flynn	702.00	Spain	Male	44.00	9.00	0.00	1.00	0.00	59207.41	1.00
9983	Ch'en	479.00	Spain	Male	35.00	4.00	125920.98	1.00	1.00	20393.44	0.00
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00

1385 rows × 11 columns

churn[(churn['gender']=='Male')&(churn['geography']=='Spain')|(churn['surname']=='Lorenzo')]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
12	Andrews	497.00	Spain	Male	24.00	3.00	0.00	2.00	0.00	76390.01	0.00
25	Muldrow	587.00	Spain	Male	45.00	6.00	0.00	1.00	0.00	158684.81	0.00
43	Watson	490.00	Spain	Male	31.00	3.00	145260.23	1.00	1.00	114066.77	0.00
44	Lorenzo	804.00	Spain	Male	33.00	7.00	76548.60	1.00	1.00	98453.45	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9976	Graham	764.00	Spain	Male	38.00	4.00	113607.47	1.00	0.00	91094.46	0.00
9978	Aldridge	573.00	Spain	Male	44.00	9.00	0.00	2.00	0.00	107124.17	0.00
9979	Flynn	702.00	Spain	Male	44.00	9.00	0.00	1.00	0.00	59207.41	1.00
9983	Ch'en	479.00	Spain	Male	35.00	4.00	125920.98	1.00	1.00	20393.44	0.00
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00

1395 rows × 11 columns

NUMBEROFPRODUCTS

QUESTION 18 WHO HAS THE HIGHEST NUMBER OF PRODUCTS

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn.numofproducts.value_counts()

1.00    5069
2.00    4578
3.00     266
4.00      60
Name: numofproducts, dtype: int64

churn.numofproducts.unique()

array([1., 3., 2., 4.])

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

QUESTION 19 , WHICH GENDER HAS THE HIGHEST AVERAGE BALANCE

churn.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

churn.groupby('gender')['balance'].mean().sort_values(ascending=False)

gender
Male     77170.09
Female   75693.28
Name: balance, dtype: float64

sns.barplot(x='gender',y='balance',data=churn)
plt.title('gender by balance')
plt.show()

Q 20,DISPLAY TO 10 BALANCE AND GEOGRAPHY

churn.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

churn_top10=churn.nlargest(10,('balance'))[['geography']]

churn_top10

	geography
2099	Spain
3287	France
8750	Spain
3595	Spain
6734	Spain
1074	Germany
145	Spain
1540	France
3927	Germany
2716	France

churn_top10=churn.nlargest(10,('balance'))[['geography']].set_index('geography')

churn_top10

geography
Spain
France
Spain
Spain
Spain
Germany
Spain
France
Germany
France

churn_top10=churn.nsmallest(10,('balance'))[['geography']]

churn_top10

	geography
0	France
3	France
12	Spain
13	France
14	France
18	Spain
24	Spain
25	Spain
26	France
27	France

churn_top10=churn.nlargest(10,('balance'))[['geography','gender']]

churn_top10

	geography	gender
2099	Spain	Male
3287	France	Female
8750	Spain	Male
3595	Spain	Male
6734	Spain	Female
1074	Germany	Male
145	Spain	Female
1540	France	Female
3927	Germany	Male
2716	France	Male

Q21 , DISPLAY TOTAL NUMBER OF SALARY IN THE DATASET

churn['estimatedsalary'].value_counts()

62825.03     3
24924.92     2
101348.88    1
182692.80    1
38466.39     1
            ..
2465.80      1
47799.15     1
165562.84    1
72927.68     1
38190.78     1
Name: estimatedsalary, Length: 9970, dtype: int64

sns.countplot(x='estimatedsalary',data=churn)
plt.title('no of estimatedsalary')
plt.show()

QUESTION 22, FIND THE MOST POPULAR GEOGRAPHY WITH HIGHEST CREDITSCORE

churn.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

churn['creditscore'].max()

850.0

churn['creditscore']

0       619.00
1       608.00
2       502.00
3       699.00
5       645.00
         ...  
9997    741.00
10006   659.00
10025   771.00
10030   772.00
10031   792.00
Name: creditscore, Length: 9973, dtype: float64

[churn['creditscore'].max()==churn['creditscore']]

[0        False
 1        False
 2        False
 3        False
 5        False
          ...  
 9997     False
 10006    False
 10025    False
 10030    False
 10031    False
 Name: creditscore, Length: 9973, dtype: bool]

churn[churn['creditscore'].max()==churn['creditscore']]

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
45	Armstrong	850.00	France	Male	36.00	7.00	0.00	1.00	1.00	40812.90	0.00
187	Rozier	850.00	Spain	Female	45.00	2.00	122311.21	1.00	1.00	19482.50	0.00
207	Chiemezie	850.00	Spain	Male	30.00	2.00	141040.01	1.00	1.00	5978.20	0.00
230	Stevenson	850.00	France	Male	33.00	10.00	0.00	1.00	0.00	4861.72	1.00
266	Welch	850.00	Germany	Male	38.00	3.00	54901.01	1.00	1.00	140075.55	0.00
...	...	...	...	...	...	...	...	...	...	...	...
9598	Glover	850.00	France	Female	28.00	9.00	0.00	2.00	1.00	164864.67	0.00
9663	Muir	850.00	Spain	Male	71.00	10.00	69608.14	1.00	0.00	97893.40	1.00
9705	Ward	850.00	France	Male	68.00	5.00	169445.40	1.00	1.00	186335.07	0.00
9948	Greco	850.00	France	Female	34.00	6.00	101266.51	1.00	0.00	33501.98	0.00
9966	Wan	850.00	France	Female	47.00	9.00	137301.87	1.00	0.00	44351.77	0.00

231 rows × 11 columns

churn[churn['creditscore'].max()==churn['creditscore']]['geography']

45       France
187       Spain
207       Spain
230      France
266     Germany
         ...   
9598     France
9663      Spain
9705     France
9948     France
9966     France
Name: geography, Length: 231, dtype: object

Q 23, DOES SALARY AFFECT BALANCE

churn.columns

Index(['surname', 'creditscore', 'geography', 'gender', 'age', 'tenure',
       'balance', 'numofproducts', 'isactivemember', 'estimatedsalary',
       'exited'],
      dtype='object')

sns.scatterplot(x='estimatedsalary',y='balance',data=churn)

<AxesSubplot:xlabel='estimatedsalary', ylabel='balance'>

Q24, CLASSIFY CREDITSCORE BASED ON EXITED

churn

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
3	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
...	...	...	...	...	...	...	...	...	...	...	...
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00
10006	Nepean	659.00	France	Male	36.00	6.00	123841.49	2.00	0.00	96833.00	0.00
10025	Obijiaku	771.00	France	Male	39.00	5.00	0.00	2.00	0.00	96270.64	0.00
10030	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	0.00	92888.52	1.00
10031	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	0.00	38190.78	0.00

9973 rows × 11 columns

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

VISUALYZING AND EXPLORATORY DATA ANALYSIS (EDA)

GEOGRAPHY

churn.geography.unique()

array(['France', 'Spain', 'Germany'], dtype=object)

sns.set(style='whitegrid')
plt.figure(figsize=(14, 7))
labels=['France', 'Spain','Germany'],
plt.pie(churn['geography'].value_counts(),labels=['France', 'Spain','Germany'],explode=[0.1,0.1,0.1],
        autopct='%1.2f%%',colors=['#E37383','#FFC0CB'], startangle=90)
plt.title('geography')
plt.axis('equal')
plt.show()

GENDER

plt.figure(figsize=(8,5))
sns.countplot('gender', data = churn, color='#00ddff', saturation=0.9)

<AxesSubplot:xlabel='gender', ylabel='count'>

plt.figure(figsize=(20,20))
sns.heatmap(churn.corr(), vmin=-1, cmap="plasma_r", annot=True)
#same thing can be seen from the correlation as well

<AxesSubplot:>

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

sns.scatterplot(x='age',y='exited',data=churn)

<AxesSubplot:xlabel='age', ylabel='exited'>

sns.scatterplot(x='age',y='exited',data=churn,hue='gender',palette='RdYlBu')

<AxesSubplot:xlabel='age', ylabel='exited'>

sns.barplot(x='gender',y='exited',data=churn,palette='RdYlBu')

<AxesSubplot:xlabel='gender', ylabel='exited'>

sns.histplot(churn['tenure'],kde=True,bins=15)

<AxesSubplot:xlabel='tenure', ylabel='Count'>

sns.boxplot(x='geography',y='exited',data=churn,hue='gender',palette='YlGnBu')

<AxesSubplot:xlabel='geography', ylabel='exited'>

sns.histplot(churn['estimatedsalary'],kde=True,bins=15)

<AxesSubplot:xlabel='estimatedsalary', ylabel='Count'>

sns.barplot(x='gender',y='estimatedsalary',data=churn,palette='RdYlBu')

<AxesSubplot:xlabel='gender', ylabel='estimatedsalary'>

sns.boxplot(x='geography',y='estimatedsalary',data=churn,hue='gender',palette='YlGnBu')

<AxesSubplot:xlabel='geography', ylabel='estimatedsalary'>

sns.histplot(churn['balance'],kde=True,bins=15)

<AxesSubplot:xlabel='balance', ylabel='Count'>

sns.barplot(x='gender',y='balance',data=churn,palette='RdYlBu')

<AxesSubplot:xlabel='gender', ylabel='balance'>

sns.boxplot(x='geography',y='balance',data=churn,hue='gender',palette='YlGnBu')

<AxesSubplot:xlabel='geography', ylabel='balance'>

sns.histplot(churn['creditscore'],kde=True,bins=15)

<AxesSubplot:xlabel='creditscore', ylabel='Count'>

sns.barplot(x='gender',y='creditscore',data=churn,palette='RdYlBu')

<AxesSubplot:xlabel='gender', ylabel='creditscore'>

sns.boxplot(x='geography',y='creditscore',data=churn,hue='gender',palette='YlGnBu')

<AxesSubplot:xlabel='geography', ylabel='creditscore'>

sns.stripplot(x='geography',y='creditscore',data=churn,hue='gender',dodge=True ,palette='YlGnBu')

<AxesSubplot:xlabel='geography', ylabel='creditscore'>

sns.stripplot(x='geography',y='exited',data=churn,hue='gender',dodge=True ,palette='YlGnBu')

<AxesSubplot:xlabel='geography', ylabel='exited'>

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

sns.jointplot(x='exited',y='balance',data=churn,kind='reg')

<seaborn.axisgrid.JointGrid at 0x1b50039c3a0>

sns.jointplot(x='exited',y='estimatedsalary',data=churn)

<seaborn.axisgrid.JointGrid at 0x1b50016b490>

sns.jointplot(x='creditscore',y='estimatedsalary',data=churn,)

<seaborn.axisgrid.JointGrid at 0x1b57ff08a00>

sns.jointplot(x='creditscore',y='estimatedsalary',data=churn,kind='kde')

<seaborn.axisgrid.JointGrid at 0x1b57ed06a00>

sns.jointplot(x='creditscore',y='estimatedsalary',data=churn,kind='kde',shade=True)

<seaborn.axisgrid.JointGrid at 0x1b57eb51310>

sns.jointplot(x='creditscore',y='estimatedsalary',data=churn,kind='kde',shade=True,cmap='YlGnBu')

<seaborn.axisgrid.JointGrid at 0x1b57e98f940>

sns.jointplot(x='creditscore',y='estimatedsalary',data=churn,kind='hex',cmap='YlGnBu')

<seaborn.axisgrid.JointGrid at 0x1b57e7cdaf0>

sns.countplot(x='exited',hue='gender',data=churn)

<AxesSubplot:xlabel='exited', ylabel='count'>

sns.countplot(x='exited',hue='geography',data=churn)

<AxesSubplot:xlabel='exited', ylabel='count'>

churn['age'].plot.hist()

<AxesSubplot:ylabel='Frequency'>

churn['balance'].plot.hist(bins=20,figsize=(10,5))

<AxesSubplot:ylabel='Frequency'>

churn['creditscore'].plot.hist(bins=20,figsize=(10,5))

<AxesSubplot:ylabel='Frequency'>

churn['estimatedsalary'].plot.hist(bins=20,figsize=(10,5))

<AxesSubplot:ylabel='Frequency'>

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn['exited'].unique

<bound method Series.unique of 0       1.00
1       0.00
2       1.00
3       0.00
5       1.00
        ... 
9997    0.00
10006   0.00
10025   0.00
10030   1.00
10031   0.00
Name: exited, Length: 9973, dtype: float64>

#visualization of this data
cols=['creditscore','geography','gender','age','tenure','balance','numofproducts','isactivemember','estimatedsalary','exited']
n_rows=2
n_cols=3
#the subplot grid and figsize of the graph
fig,  axs =plt.subplots(n_rows,n_cols,figsize=(n_cols*3.2,n_rows*3.2))
for r in range(0,n_rows):
    for c in range (0,n_cols):
        i = r*n_cols+c  #index to go through the number of columns
        ax= axs[r][c]   #show were to position each subplot
        sns.countplot(churn[cols[i]],hue=churn['exited'],ax=ax)
        ax.set_title(cols[i])
        ax.legend(title='exited',loc='upper right')
        plt.tight_layout

churn.pivot_table('exited',index='gender',columns='geography').plot()

<AxesSubplot:xlabel='gender'>

f, ax =plt.subplots(1,2,figsize= (18,8))
churn['exited'].value_counts().plot.pie(explode=[0,0.1,],autopct='%1.1f%%',ax=ax[0],shadow=True)
ax[0].set_title('exited')
ax[0].set_ylabel('')
sns.countplot('exited',data=churn,ax=ax[1])
ax[1].set_title('exited')
plt.show()

churn.head(2)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00

churn_gender = churn.groupby(['gender'])[['creditscore', 'balance', 'estimatedsalary']].mean()

colors_4 = ['magenta','yellow','green','red']
colors_3 =['magenta','yellow','green']

churn_gender.plot.pie(subplots=True, figsize=(20,10), labels=churn_gender.index, autopct='%1.1f%%',colors=colors_4)
plt.show()

churn_geography = churn.groupby(['geography'])[['creditscore', 'balance', 'estimatedsalary']].mean()

churn_geography.plot.pie(subplots=True, figsize=(20,10), labels=churn_geography.index, autopct='%1.1f%%',colors=colors_4)
plt.show()

sns.factorplot('geography','exited',hue='gender',data=churn)

<seaborn.axisgrid.FacetGrid at 0x1b51c537a30>

f, ax = plt.subplots(1,2,figsize=(18,8,))
sns.violinplot('geography','age', hue='exited',data=churn,ax=ax[0])
ax[0].set_title('geography and age vs exited')
ax[0].set_yticks(range(0,110,10))
sns.violinplot('gender','age',hue='exited',data=churn,ax=ax[1])
ax[1].set_title('gender and age vs exited')
ax[1].set_yticks(range(0,110,10))
plt.show()

f, ax=plt.subplots(1,2,figsize=(12,10))
churn[churn['exited']==0].age.plot.hist(ax=ax[0],bins=20,edgecolor='black',color='red')
ax[0].set_title('exited=0')
x1=list(range(0,85,5))
ax[0].set_xticks(x1)
churn[churn['exited']==1].age.plot.hist(ax=ax[1],color='green',bins=20,edgecolor='black')
ax[1].set_title('exited=1')
x2=list(range(0,85,5))
ax[1].set_xticks(x2)
plt.show()

plt.figure(figsize=(30,30))
sns.set_context("talk",font_scale=1)
sns.set_palette("pastel")
ax = sns.countplot(y="age", hue="gender", data=churn)
ax.legend(loc='upper right',frameon=True)
plt.title('GENDER vs AGE', fontsize=18, fontweight='bold')
ax.set(xlabel='COUNT OF GENDER',ylabel='AGE')
plt.show()

plt.figure(figsize=(6,6))
sns.set_context("talk",font_scale=1)
sns.set_palette("pastel")
ax = sns.countplot(y="gender", hue="exited", data=churn)
ax.legend(loc='upper right',frameon=True)
plt.title('exited vs gender', fontsize=8, fontweight='bold')
ax.set(xlabel='COUNT OF exited ',ylabel='gender')
plt.show()

plt.figure(figsize=(8,8))
plt.title('Correlation Analysis',color='Red',fontsize=20,pad=40)

corr = churn.corr()
mask = np.triu(np.ones_like(corr,dtype = bool))
sns.heatmap(churn.corr(),mask=mask,annot=True,linewidths=.5);
plt.xticks(rotation=60)
plt.yticks(rotation = 60)
plt.show()

colors = ['#E94B3C','#2D2926']

exited = churn[churn['exited'] == 1].describe().T
not_exited = churn[churn['exited'] == 0].describe().T

fig,ax = plt.subplots(nrows = 1,ncols = 2,figsize = (5,5))
plt.subplot(1,2,1)
sns.heatmap(exited[['mean']],annot = True,cmap = colors,linewidths = 0.4,linecolor = 'black',cbar = False,fmt = '.2f')
plt.title('exited Customers');

plt.subplot(1,2,2)
sns.heatmap(not_exited[['mean']],annot = True,cmap = colors,linewidths = 0.4,linecolor = 'black',cbar = False,fmt = '.2f',)
plt.title('Not_exited Customers');

fig.tight_layout(pad = 0)

l = list(churn['exited'].value_counts())
circle = [l[0] / sum(l) * 100,l[1] / sum(l) * 100]

fig = plt.subplots(nrows = 1,ncols = 2,figsize = (20,5))
plt.subplot(1,2,1)
plt.pie(circle,labels = ['Not-exited Customer','exited Customer'],autopct = '%1.1f%%',startangle = 90,explode = (0.1,0),colors = colors,
       wedgeprops = {'edgecolor' : 'black','linewidth': 1,'antialiased' : True})
plt.title('exited - Not-exited %');

plt.subplot(1,2,2)
ax = sns.countplot('exited',data = churn,palette = colors,edgecolor = 'black')
for rect in ax.patches:
    ax.text(rect.get_x() + rect.get_width() / 2, rect.get_height() + 2, rect.get_height(), horizontalalignment='center', fontsize = 11)

    
plt.title('Number of Exited - Not-Exited Customers');
plt.show()

DATA SCIENCE

churn.head(5)

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
3	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00

DATA PREPROCESSING

converting categorical variable to numerical cvariable we can

use 2 methods ENCODING AND MAPP METHODS

MAP METHOD OF CONVERTING CATEGORICAL TO NUMERICAL

churn.head()

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
3	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00

churn=churn.drop(churn.columns[[0]],axis=1)

churn.dtypes

creditscore        float64
geography           object
gender              object
age                float64
tenure             float64
balance            float64
numofproducts      float64
isactivemember     float64
estimatedsalary    float64
exited             float64
dtype: object

churn_catg

	surname	geography	gender
0	Hargrave	France	Female
1	Hill	Spain	Female
2	Onio	France	Female
3	Boni	France	Female
4	NaN	NaN	NaN
...	...	...	...
10027	Johnstone	France	Male
10028	Liu	France	Female
10029	Liu	France	Female
10030	Sabbatini	Germany	Male
10031	Walker	France	Female

10032 rows × 3 columns

churn_cont.head(2)

	rownumber	customerid	creditscore	age	tenure	balance	numofproducts	hascrcard	isactivemember	estimatedsalary	exited
0	1.00	15634602.00	619.00	42.00	2.00	0.00	1.00	1.00	1.00	101348.88	1.00
1	2.00	NaN	608.00	41.00	1.00	83807.86	1.00	0.00	1.00	112542.58	0.00

churn1

	surname	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	Hargrave	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	Hill	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	Onio	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
3	Boni	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	Chu	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00
...	...	...	...	...	...	...	...	...	...	...	...
9997	T'ao	741.00	Spain	Male	35.00	6.00	74371.49	1.00	0.00	99595.67	0.00
10006	Nepean	659.00	France	Male	36.00	6.00	123841.49	2.00	0.00	96833.00	0.00
10025	Obijiaku	771.00	France	Male	39.00	5.00	0.00	2.00	0.00	96270.64	0.00
10030	Sabbatini	772.00	Germany	Male	42.00	3.00	75075.31	2.00	0.00	92888.52	1.00
10031	Walker	792.00	France	Female	28.00	4.00	130142.79	1.00	0.00	38190.78	0.00

9913 rows × 11 columns

churn.head()

	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
3	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00

REPLACE METHOD

#geography1={'geography':{'France':0,'Spain':1,'Germany':2}}

#churn=churn.replace(geography1)

churn.head()

	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited
0	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00
1	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00
2	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00
3	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00
5	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00

LABEL ENCODER METHOD

from sklearn.preprocessing import LabelEncoder

le=LabelEncoder()

churn['geography1']=le.fit_transform(churn.geography)

churn.head()

	creditscore	geography	gender	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1
0	619.00	France	Female	42.00	2.00	0.00	1.00	1.00	101348.88	1.00	0
1	608.00	Spain	Female	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00	2
2	502.00	France	Female	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00	0
3	699.00	France	Female	39.00	1.00	0.00	2.00	0.00	93826.63	0.00	0
5	645.00	Spain	Male	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00	2

dummies variable methiod

churn=pd.get_dummies(churn,drop_first=True)

churn.head()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	1.00	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0.00	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00	2	0	1	1

#churn=pd.get_dummies(churn,columns=['gender'],prefix='gender',drop_first=True)

INBALANCE DATASET

UNDERSAMPLING AND OVERSAMPLING METHODS

churn.head(2)

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	1.00	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00	2	0	1	0

churn['exited'].value_counts()

0.00    7943
1.00    2030
Name: exited, dtype: int64

sns.countplot(churn['exited'])

<AxesSubplot:xlabel='exited', ylabel='count'>

store feature matrix in X and response target in vector y

X= churn.drop('exited',axis=1)

y=churn['exited']

X

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	2	0	1	1
...	...	...	...	...	...	...	...	...	...	...	...
9997	741.00	35.00	6.00	74371.49	1.00	0.00	99595.67	2	0	1	1
10006	659.00	36.00	6.00	123841.49	2.00	0.00	96833.00	0	0	0	1
10025	771.00	39.00	5.00	0.00	2.00	0.00	96270.64	0	0	0	1
10030	772.00	42.00	3.00	75075.31	2.00	0.00	92888.52	1	1	0	1
10031	792.00	28.00	4.00	130142.79	1.00	0.00	38190.78	0	0	0	0

9973 rows × 11 columns

y

0       1.00
1       0.00
2       1.00
3       0.00
5       1.00
        ... 
9997    0.00
10006   0.00
10025   0.00
10030   1.00
10031   0.00
Name: exited, Length: 9973, dtype: float64

split dataset into training set and test set

from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.20,random_state=42,stratify=y)

X_train

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	geography1	geography_Germany	geography_Spain	gender_Male
5800	624.00	71.00	7.00	0.00	2.00	1.00	108841.83	0	0	0	1
5401	727.00	28.00	1.00	0.00	1.00	0.00	40357.39	2	0	1	1
2369	709.00	45.00	4.00	122917.71	1.00	1.00	11.58	1	1	0	1
7532	623.00	35.00	0.00	130557.24	1.00	1.00	47880.71	0	0	0	0
3495	587.00	35.00	3.00	83286.56	1.00	0.00	125553.52	2	0	1	0
...	...	...	...	...	...	...	...	...	...	...	...
9692	735.00	39.00	1.00	60374.98	1.00	0.00	40223.74	2	0	1	1
4820	517.00	39.00	3.00	0.00	2.00	1.00	12465.51	2	0	1	1
3082	556.00	35.00	10.00	0.00	2.00	1.00	192751.18	0	0	0	1
2558	537.00	53.00	3.00	0.00	1.00	1.00	91406.62	0	0	0	0
9301	633.00	61.00	3.00	157201.48	1.00	1.00	50368.63	0	0	0	1

7978 rows × 11 columns

y_train

5800   0.00
5401   0.00
2369   1.00
7532   0.00
3495   0.00
       ... 
9692   0.00
4820   0.00
3082   0.00
2558   0.00
9301   0.00
Name: exited, Length: 7978, dtype: float64

MACHINE LEARNING ALGRITHMS

LOGISTIC REGRESSION

churn.head()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	1.00	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0.00	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00	2	0	1	1

X

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	2	0	1	1
...	...	...	...	...	...	...	...	...	...	...	...
9997	741.00	35.00	6.00	74371.49	1.00	0.00	99595.67	2	0	1	1
10006	659.00	36.00	6.00	123841.49	2.00	0.00	96833.00	0	0	0	1
10025	771.00	39.00	5.00	0.00	2.00	0.00	96270.64	0	0	0	1
10030	772.00	42.00	3.00	75075.31	2.00	0.00	92888.52	1	1	0	1
10031	792.00	28.00	4.00	130142.79	1.00	0.00	38190.78	0	0	0	0

9973 rows × 11 columns

FEAUTURE SCALING

from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler

sc=StandardScaler()

X_train=sc.fit_transform(X_train)

X_test=sc.fit_transform(X_test)

X_train

array([[-0.26906411,  3.04660636,  0.69093269, ..., -0.57918359,
        -0.57628887,  0.9139413 ],
       [ 0.79139897, -1.03510395, -1.38121605, ..., -0.57918359,
         1.73524086,  0.9139413 ],
       [ 0.60607532,  0.57859547, -0.34514168, ...,  1.72656825,
        -0.57628887,  0.9139413 ],
       ...,
       [-0.96917566, -0.37063949,  1.72700706, ..., -0.57918359,
        -0.57628887,  0.9139413 ],
       [-1.16479506,  1.33798344, -0.6904998 , ..., -0.57918359,
        -0.57628887, -1.09416217],
       [-0.17640229,  2.0973714 , -0.6904998 , ..., -0.57918359,
        -0.57628887,  0.9139413 ]])

from sklearn.linear_model import LogisticRegression

log=LogisticRegression()

log.fit(X_train,y_train)

LogisticRegression()

y_pred1=log.predict(X_test)

from sklearn.metrics import accuracy_score

accuracy_score(y_test,y_pred1)

0.8095238095238095

from sklearn.metrics import precision_score,recall_score,f1_score

precision_score(y_test,y_pred1)

0.5955882352941176

recall_score(y_test,y_pred1)

0.19950738916256158

f1_score(y_test,y_pred1)

0.2988929889298893

EVALUATE IMBALANCE DATA SET FIRST BY UNDERSAMPLING

normal=churn[churn['exited']==0]
fraud=churn[churn['exited']==1]

normal.shape

(7943, 12)

fraud.shape

(2030, 12)

normal_sample=normal.sample(n=2055)

normal_sample.shape

(2055, 12)

new_churn=pd.concat([normal_sample,fraud])

new_churn['exited'].value_counts()

0.00    2055
1.00    2030
Name: exited, dtype: int64

new_churn.head()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1	geography_Germany	geography_Spain	gender_Male
3921	658.00	44.00	2.00	168396.34	1.00	1.00	14178.73	0.00	0	0	0	1
8311	704.00	36.00	2.00	175509.80	2.00	0.00	152039.67	0.00	2	0	1	0
5438	589.00	39.00	7.00	0.00	2.00	0.00	95985.64	0.00	0	0	0	0
3044	693.00	21.00	1.00	0.00	2.00	1.00	3494.02	0.00	0	0	0	1
8989	650.00	32.00	4.00	79450.09	1.00	1.00	118324.75	0.00	2	0	1	0

X=new_churn.drop('exited',axis=1)

y=new_churn['exited']

from sklearn.model_selection import train_test_split
X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.20,random_state=42)

log=LogisticRegression()
log.fit(X_train,y_train)

LogisticRegression()

y_pred1=log.predict(X_test)

accuracy_score(y_test,y_pred1)

0.594859241126071

from sklearn.metrics import precision_score,recall_score,f1_score

precision_score(y_test,y_pred1)

0.5618448637316562

recall_score(y_test,y_pred1)

0.6871794871794872

f1_score(y_test,y_pred1)

0.6182237600922722

DECISION TREE CLASSIFIER

from sklearn.tree import DecisionTreeClassifier

dt=DecisionTreeClassifier()

dt.fit(X_train,y_train)

DecisionTreeClassifier()

y_pred2=dt.predict(X_test)

accuracy_score(y_test,y_pred2)

0.7025703794369645

f1_score(y_test,y_pred2)

0.6864516129032259

precision_score(y_test,y_pred2)

0.6909090909090909

recall_score(y_test,y_pred2)

0.6820512820512821

RANDOM FOREST CLASSIFIER

from sklearn.ensemble import RandomForestClassifier

rf=RandomForestClassifier()

rf.fit(X_train,y_train)

RandomForestClassifier()

y_pred3=rf.predict(X_test)

accuracy_score(y_test,y_pred3)

0.7870257037943696

f1_score(y_test,y_pred3)

0.7740259740259741

precision_score(y_test,y_pred3)

0.7842105263157895

recall_score(y_test,y_pred3)

0.764102564102564

visualizing the results after after undersampling

final_data=pd.DataFrame({'Models':['log','dt','rf'],
             'ACC':[accuracy_score(y_test,y_pred1)*100,
                                               
                                                    
                    accuracy_score(y_test,y_pred2)*100,
                    accuracy_score(y_test,y_pred3)*100]}) 

final_data

	Models	ACC
0	log	59.49
1	dt	70.26
2	rf	78.70

sns.barplot(final_data['Models'],final_data['ACC'])

<AxesSubplot:xlabel='Models', ylabel='ACC'>

OVERSAMPLING USING SMOTE

synthetic minority over sampling

churn.head()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	1.00	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0.00	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00	2	0	1	1

churn.shape

(9973, 12)

churn['exited'].value_counts()

0.00    7943
1.00    2030
Name: exited, dtype: int64

sns.countplot(churn['exited'])

<AxesSubplot:xlabel='exited', ylabel='count'>

X= churn.drop('exited',axis=1)

X

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	2	0	1	1
...	...	...	...	...	...	...	...	...	...	...	...
9997	741.00	35.00	6.00	74371.49	1.00	0.00	99595.67	2	0	1	1
10006	659.00	36.00	6.00	123841.49	2.00	0.00	96833.00	0	0	0	1
10025	771.00	39.00	5.00	0.00	2.00	0.00	96270.64	0	0	0	1
10030	772.00	42.00	3.00	75075.31	2.00	0.00	92888.52	1	1	0	1
10031	792.00	28.00	4.00	130142.79	1.00	0.00	38190.78	0	0	0	0

9973 rows × 11 columns

y=churn['exited']

y

0       1.00
1       0.00
2       1.00
3       0.00
5       1.00
        ... 
9997    0.00
10006   0.00
10025   0.00
10030   1.00
10031   0.00
Name: exited, Length: 9973, dtype: float64

from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.20,random_state=42,stratify=y)

X_train

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	geography1	geography_Germany	geography_Spain	gender_Male
5800	624.00	71.00	7.00	0.00	2.00	1.00	108841.83	0	0	0	1
5401	727.00	28.00	1.00	0.00	1.00	0.00	40357.39	2	0	1	1
2369	709.00	45.00	4.00	122917.71	1.00	1.00	11.58	1	1	0	1
7532	623.00	35.00	0.00	130557.24	1.00	1.00	47880.71	0	0	0	0
3495	587.00	35.00	3.00	83286.56	1.00	0.00	125553.52	2	0	1	0
...	...	...	...	...	...	...	...	...	...	...	...
9692	735.00	39.00	1.00	60374.98	1.00	0.00	40223.74	2	0	1	1
4820	517.00	39.00	3.00	0.00	2.00	1.00	12465.51	2	0	1	1
3082	556.00	35.00	10.00	0.00	2.00	1.00	192751.18	0	0	0	1
2558	537.00	53.00	3.00	0.00	1.00	1.00	91406.62	0	0	0	0
9301	633.00	61.00	3.00	157201.48	1.00	1.00	50368.63	0	0	0	1

7978 rows × 11 columns

y_train

5800   0.00
5401   0.00
2369   1.00
7532   0.00
3495   0.00
       ... 
9692   0.00
4820   0.00
3082   0.00
2558   0.00
9301   0.00
Name: exited, Length: 7978, dtype: float64

from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import MinMaxScaler

minmax = MinMaxScaler()

X_train=minmax.fit_transform(X_train)

X_test=minmax.fit_transform(X_test)

X_train

array([[0.548     , 0.71621622, 0.7       , ..., 0.        , 0.        ,
        1.        ],
       [0.754     , 0.13513514, 0.1       , ..., 0.        , 1.        ,
        1.        ],
       [0.718     , 0.36486486, 0.4       , ..., 1.        , 0.        ,
        1.        ],
       ...,
       [0.412     , 0.22972973, 1.        , ..., 0.        , 0.        ,
        1.        ],
       [0.374     , 0.47297297, 0.3       , ..., 0.        , 0.        ,
        0.        ],
       [0.566     , 0.58108108, 0.3       , ..., 0.        , 0.        ,
        1.        ]])

from imblearn.over_sampling  import SMOTE

SMOTE().fit_resample(X,y)

(       creditscore   age  tenure   balance  numofproducts  isactivemember  \
 0           619.00 42.00    2.00      0.00           1.00            1.00   
 1           608.00 41.00    1.00  83807.86           1.00            1.00   
 2           502.00 42.00    8.00 159660.80           3.00            0.00   
 3           699.00 39.00    1.00      0.00           2.00            0.00   
 4           645.00 44.00    8.00 113755.78           2.00            0.00   
 ...            ...   ...     ...       ...            ...             ...   
 15881       757.07 60.50    7.98 125438.21           1.05            0.00   
 15882       579.83 49.40    2.35 100032.41           1.00            0.41   
 15883       601.86 43.62    2.57      0.00           1.57            0.57   
 15884       571.98 39.63    4.00      0.00           1.00            0.00   
 15885       607.70 45.60    4.02      0.00           1.49            0.00   
 
        estimatedsalary  geography1  geography_Germany  geography_Spain  \
 0            101348.88           0                  0                0   
 1            112542.58           2                  0                1   
 2            113931.57           0                  0                0   
 3             93826.63           0                  0                0   
 4            149756.71           2                  0                1   
 ...                ...         ...                ...              ...   
 15881        182223.25           0                  0                0   
 15882          5928.03           1                  0                0   
 15883        174137.68           0                  0                0   
 15884         95434.59           0                  0                0   
 15885         92236.98           1                  0                0   
 
        gender_Male  
 0                0  
 1                0  
 2                0  
 3                0  
 4                1  
 ...            ...  
 15881            0  
 15882            1  
 15883            0  
 15884            0  
 15885            0  
 
 [15886 rows x 11 columns],
 0       1.00
 1       0.00
 2       1.00
 3       0.00
 4       1.00
         ... 
 15881   1.00
 15882   1.00
 15883   1.00
 15884   1.00
 15885   1.00
 Name: exited, Length: 15886, dtype: float64)

X_res,y_res=SMOTE().fit_resample(X,y)

y_res.value_counts()

1.00    7943
0.00    7943
Name: exited, dtype: int64

X_train,X_test,y_train,y_test=train_test_split(X_res,y_res,test_size=0.20,random_state=42,)

from sklearn.linear_model import LogisticRegression

log=LogisticRegression()

log.fit(X_train,y_train)

LogisticRegression()

y_pred1=log.predict(X_test)

from sklearn.metrics import accuracy_score

accuracy_score(y_test,y_pred1)

0.6784140969162996

from sklearn.metrics import precision_score,recall_score,f1_score

precision_score(y_test,y_pred1)

0.6680722891566265

recall_score(y_test,y_pred1)

0.7018987341772152

f1_score(y_test,y_pred1)

0.6845679012345679

SVC

from sklearn import svm

svm=svm.SVC()

svm.fit(X_train,y_train)

SVC()

y_pred2=svm.predict(X_test)

accuracy_score(y_test,y_pred2)

0.5670232850849591

precision_score(y_test,y_pred2)

0.5473098330241187

f1_score(y_test,y_pred2)

0.6316916488222698

KNeighbors Classifier

from sklearn.neighbors import KNeighborsClassifier

knn=KNeighborsClassifier()

knn.fit(X_train,y_train)

KNeighborsClassifier()

y_pred3=knn.predict(X_test)

accuracy_score(y_test,y_pred3)

0.6768407803650094

precision_score(y_test,y_pred3)

0.6469962785752259

recall_score(y_test,y_pred3)

0.770253164556962

f1_score(y_test,y_pred3)

0.7032649523259173

Decision Tree Classifier

from sklearn.tree import DecisionTreeClassifier

dt=DecisionTreeClassifier()

dt.fit(X_train,y_train)

DecisionTreeClassifier()

y_pred4=dt.predict(X_test)

accuracy_score(y_test,y_pred4)

0.8442416614222782

precision_score(y_test,y_pred4)

0.84098051539912

recall_score(y_test,y_pred4)

0.8468354430379746

f1_score(y_test,y_pred4)

0.8438978240302742

Random Forest Classifier

from sklearn.ensemble import RandomForestClassifier

rf=RandomForestClassifier()

rf.fit(X_train,y_train)

RandomForestClassifier()

y_pred5=rf.predict(X_test)

accuracy_score(y_test,y_pred5)

0.89490245437382

precision_score(y_test,y_pred5)

0.9192462987886945

recall_score(y_test,y_pred5)

0.8645569620253165

f1_score(y_test,y_pred5)

0.8910632746249184

Gradient Boosting Classifier

from sklearn.ensemble import GradientBoostingClassifier

gb=GradientBoostingClassifier()

gb.fit(X_train,y_train)

GradientBoostingClassifier()

y_pred6=gb.predict(X_test)

accuracy_score(y_test,y_pred6)

0.8926998112020138

precision_score(y_test,y_pred6)

0.9166106254203094

recall_score(y_test,y_pred6)

0.8626582278481013

f1_score(y_test,y_pred6)

0.8888164329964134

SAVE THE MODEL

final_data1=pd.DataFrame({'Models':['log','svc','knn','dt','rf','gb'],
             'ACC':[accuracy_score(y_test,y_pred1)*100,
                    accuracy_score(y_test,y_pred2)*100,                           
                    accuracy_score(y_test,y_pred3)*100,                                
                    accuracy_score(y_test,y_pred4)*100,
                    accuracy_score(y_test,y_pred5)*100,
                    accuracy_score(y_test,y_pred6)*100]}) 

final_data1

	Models	ACC
0	log	67.84
1	svc	56.70
2	knn	67.68
3	dt	84.42
4	rf	89.49
5	gb	89.27

sns.barplot(final_data1['Models'],final_data1['ACC'])

<AxesSubplot:xlabel='Models', ylabel='ACC'>

final_data2=pd.DataFrame({'Models':['log','svc','knn','dt','rf','gb'],
             'PRE':[precision_score(y_test,y_pred1)*100,
                    precision_score(y_test,y_pred2)*100,                           
                    precision_score(y_test,y_pred3)*100,                                
                    precision_score(y_test,y_pred4)*100,
                    precision_score(y_test,y_pred5)*100,
                    precision_score(y_test,y_pred6)*100]}) 

final_data2

	Models	PRE
0	log	66.81
1	svc	54.73
2	knn	64.70
3	dt	84.10
4	rf	91.92
5	gb	91.66

sns.barplot(final_data2['Models'],final_data2['PRE'])

<AxesSubplot:xlabel='Models', ylabel='PRE'>

final_data3=pd.DataFrame({'Models':['log','svc','knn','dt','rf','gb'],
             'REC':[recall_score(y_test,y_pred1)*100,
                    recall_score(y_test,y_pred2)*100,                           
                    recall_score(y_test,y_pred3)*100,                                
                    recall_score(y_test,y_pred4)*100,
                    recall_score(y_test,y_pred5)*100,
                    recall_score(y_test,y_pred6)*100]}) 

final_data3

	Models	REC
0	log	70.19
1	svc	74.68
2	knn	77.03
3	dt	84.68
4	rf	86.46
5	gb	86.27

sns.barplot(final_data3['Models'],final_data3['REC'])

<AxesSubplot:xlabel='Models', ylabel='REC'>

final_data4=pd.DataFrame({'Models':['log','svc','knn','dt','rf','gb'],
             'F1':[f1_score(y_test,y_pred1)*100,
                    f1_score(y_test,y_pred2)*100,                           
                    f1_score(y_test,y_pred3)*100,                                
                    f1_score(y_test,y_pred4)*100,
                    f1_score(y_test,y_pred5)*100,
                    f1_score(y_test,y_pred6)*100]}) 

final_data4

	Models	F1
0	log	68.46
1	svc	63.17
2	knn	70.33
3	dt	84.39
4	rf	89.11
5	gb	88.88

sns.barplot(final_data4['Models'],final_data4['F1'])

<AxesSubplot:xlabel='Models', ylabel='F1'>

import matplotlib.pyplot as plt
%matplotlib inline
import pandas as pd

DATA SCIENCE DEVELOPER

SAVING AND DEVELOPING THE MODEL AND DEPLOYMENT

X_res=minmax.fit_transform(X_res)

rf.fit(X_res,y_res)

RandomForestClassifier()

import joblib

joblib.dump(rf,'churn_predict_model')

['churn_predict_model']

model=joblib.load('churn_predict_model')

churn.columns

Index(['creditscore', 'age', 'tenure', 'balance', 'numofproducts',
       'isactivemember', 'estimatedsalary', 'exited', 'geography1',
       'geography_Germany', 'geography_Spain', 'gender_Male'],
      dtype='object')

churn.head()

	creditscore	age	tenure	balance	numofproducts	isactivemember	estimatedsalary	exited	geography1	geography_Germany	geography_Spain	gender_Male
0	619.00	42.00	2.00	0.00	1.00	1.00	101348.88	1.00	0	0	0	0
1	608.00	41.00	1.00	83807.86	1.00	1.00	112542.58	0.00	2	0	1	0
2	502.00	42.00	8.00	159660.80	3.00	0.00	113931.57	1.00	0	0	0	0
3	699.00	39.00	1.00	0.00	2.00	0.00	93826.63	0.00	0	0	0	0
5	645.00	44.00	8.00	113755.78	2.00	0.00	149756.71	1.00	2	0	1	1

model.predict([[619,0,42,2,0.00,1.00,1.00,101348.88,1.00,0,0]])

array([0.])

GUI

from tkinter import *

import joblib

from tkinter import messagebox

master=Tk()





master.mainloop()

#ADD TITLE TO IT
master=Tk()

master.title('churn_predict_model')



master.mainloop()

#LETS SET DIMENSION OF THIS WINDOW
master=Tk()

master.title('churn_predict_model')

master.geometry('400x300')
master.minsize(200,200)
master.maxsize(600,600)
master.mainloop()

#LETS ADD COLOR TO THE WINDOW

master=Tk()

master.title('churn_predict_model')

master.geometry('400x300')
master.minsize(200,200)
master.maxsize(600,600)
master.configure(bg='lightblue')







master.mainloop()

master=Tk()
master.title('churn_predict_model')

master.geometry('400x300')
master.minsize(200,200)
master.maxsize(600,600)
master.configure(bg='lightblue')


label=Label(master,text='Bank churn',font=('Arial',20,'bold'),width=10,height=1,bg='lightblue',foreground='yellow').grid(row=0,columnspan=2)

Label(master,text='creditscore').grid(row=1)
Label(master,text='gender[0-1]').grid(row=2)
Label(master,text='age').grid(row=3)
Label(master,text='tenure').grid(row=4)
Label(master,text='balance').grid(row=5)
Label(master,text='	numofproducts').grid(row=6)
Label(master,text='isactivemember').grid(row=7)
Label(master,text='estimatedsalary').grid(row=8)
Label(master,text='geography1').grid(row=9)
Label(master,text='geography_Germany').grid(row=10)
Label(master,text='geography_Spain').grid(row=11)
master.mainloop()

master=Tk()
master.title('churn_predict_model')

master.geometry('400x300')
master.minsize(200,200)
master.maxsize(600,600)
master.configure(bg='lightblue')


label=Label(master,text='Bank churn',font=('Arial',20,'bold'),width=10,height=1,bg='lightblue',foreground='yellow').grid(row=0,columnspan=2)

Label(master,text='creditscore').grid(row=1)
Label(master,text='gender[0-1]').grid(row=2)
Label(master,text='age').grid(row=3)
Label(master,text='tenure').grid(row=4)
Label(master,text='balance').grid(row=5)
Label(master,text='	numofproducts').grid(row=6)
Label(master,text='isactivemember').grid(row=7)
Label(master,text='estimatedsalary').grid(row=8)
Label(master,text='geography1').grid(row=9)
Label(master,text='geography_Germany').grid(row=10)
Label(master,text='geography_Spain').grid(row=11)


e1=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e2=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e3=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e4=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e5=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e6=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e7=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e8=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e9=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e10=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')
e11=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3,show='*')


master.mainloop()

master=Tk()
master.title('churn_predict_model')


master.geometry('400x300')
master.minsize(200,200)
master.maxsize(600,600)
master.configure(bg='lightblue')


label=Label(master,text='Bank churn',font=('Arial',20,'bold'),width=10,height=1,bg='lightblue',foreground='yellow').grid(row=0,columnspan=2)

Label(master,text='creditscore').grid(row=1)
Label(master,text='gender[0-1]').grid(row=2)
Label(master,text='age').grid(row=3)
Label(master,text='tenure').grid(row=4)
Label(master,text='balance').grid(row=5)
Label(master,text='	numofproducts').grid(row=6)
Label(master,text='isactivemember').grid(row=7)
Label(master,text='estimatedsalary').grid(row=8)
Label(master,text='geography1').grid(row=9)
Label(master,text='geography_Germany').grid(row=10)
Label(master,text='geography_Spain').grid(row=11)


e1=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e2=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e3=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e4=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e5=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e6=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e7=Entry( master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e8=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e9=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e10=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e11=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
  

e1.grid(row=1,column=1)    
e2.grid(row=2,column=1)
e3.grid(row=3,column=1)
e4.grid(row=4,column=1)
e5.grid(row=5,column=1)
e6.grid(row=6,column=1)
e7.grid(row=7,column=1)
e8.grid(row=8,column=1)
e9.grid(row=9,column=1)
e10.grid(row=10,column=1)
e11.grid(row=11,column=1)

master.mainloop()

def show_entry():
    p1=float(e1.get())
    p2=float(e2.get())
    p3=float(e3.get())
    p4=float(e4.get())
    p5=float(e5.get())
    p6=float(e6.get())
    p7=float(e6.get())
    p8=float(e6.get())
    p9=float(e6.get())
    p10=float(e6.get())
    p11=float(e6.get())

    model=joblib.load('churn_predict_model')
    result=model.predict([[p1,p2,p3,p4,p5,p6,p7,p8,p9,p10,p11]])
    
    
    Label(master, text='Bank_churn').grid(row=12)
    Label(master,text=result).grid(row=13)
    
    
master=Tk()

master.title('churn_predict_model')


master.geometry('400x300')
master.minsize(200,200)
master.maxsize(600,600)
master.configure(bg='lightblue')


label=Label(master,text='Bank_churn',font=('Arial',20,'bold'),width=10,height=1,bg='lightblue',foreground='yellow').grid(row=0,columnspan=2)

Label(master,text='creditscore').grid(row=1)
Label(master,text='gender[0-1]').grid(row=2)
Label(master,text='age').grid(row=3)
Label(master,text='tenure').grid(row=4)
Label(master,text='balance').grid(row=5)
Label(master,text='	numofproducts').grid(row=6)
Label(master,text='isactivemember').grid(row=7)
Label(master,text='estimatedsalary').grid(row=8)
Label(master,text='geography1').grid(row=9)
Label(master,text='geography_Germany').grid(row=10)
Label(master,text='geography_Spain').grid(row=11)


e1=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e2=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e3=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e4=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e5=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e6=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e7=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)  
e8=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e9=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e10=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
e11=Entry(master,font=('Arial',14),bg='grey',fg='white',borderwidth=3)
  

e1.grid(row=1,column=1)    
e2.grid(row=2,column=1)
e3.grid(row=3,column=1)
e4.grid(row=4,column=1)
e5.grid(row=5,column=1)
e6.grid(row=6,column=1)
e7.grid(row=7,column=1)
e8.grid(row=8,column=1)
e9.grid(row=9,column=1)
e10.grid(row=10,column=1)
e11.grid(row=11,column=1)





Button(master,text='predict',command=show_entry,bg='pink',font=('Arial',20,'bold'),borderwidth=3,activebackground='blue').grid()







master.mainloop()